Development of Magnetized Plasma Sputtering Source for
Effective Target Utilization with Various Magnet Setups
September 2018
Department of Science and Advanced Technology
Graduate School of Science and Engineering
Saga University, Japan
MD. AMZAD HOSSAIN
Development of Magnetized Plasma Sputtering Source for
Effective Target Utilization with Various Magnet Setups
By
MD. AMZAD HOSSAIN
A dissertation submitted in partial fulfillment of the
requirements for the degree of
Doctor of Engineering (Dr. Eng.)
in Electrical and Electronic Engineering
Department of Science and Advanced Technology
Graduate School of Science and Engineering
Saga University, Japan
September 2018
i
Acknowledgement
First of all, Alhamdulillah, I would like to express my praise and gratitude to the Almighty,
Gracious and Merciful Allah who makes me capable to successfully complete this doctor study.
His presence is the main source of my inspiration, without his help; I would not achieve this goal.
With great pleasure, I am profoundly indebted to my supervisors Professor Dr. Yasunori
Ohtsu, Professor Dr. Toshiyuki Oishi, Professor Dr. Tooru Tanaka and Associate Professor Dr.
Satoshi Ihara of Saga University, Japan, for their expertise guidance, extraordinary kindness,
inspiration and all kinds of supports throughout my present study. Their sufficient cooperation
made me confidential enough in each step of my doctoral works.
Md. Amzad Hossain is grateful to Dr. Tatsuo Tabaru, National Institute of Advanced
Industrial Science and Technology, Professor Hirotaka Toyoda, Nagoya University, Professor Dr.
Michael A. Lieberman, University of California at Berkeley, Professor Dr. Ken Yukimura,
National Institute of Advanced Industrial Science and Technology, Japan, Dr. Julian Schulze, Ruhr
University Bochum, Germany and West Virginia University, USA for fruitful discussions.
I am thankful to Mr. Yusuke Takada, Mr. Tsubasa Ide, Mr. Koichirou Ikari, Mr. Yutaro
Nakamura, Mr. Kosei Sugawara, and Mr. Masaya Takasaki, Plasma Electronics Lab, Saga
University, Japan for their heartfelt cooperation. Warmest and sincere thanks to my entire
laboratory members who have extended their kind helps to me during my study. I would like to
expand my gratitude to the Ministry of Education, Culture, Sports, Science and Technology
(MEXT), Japan for supporting his scholarship.
I would like to thank Jessore University of Science and Technology, Jessore – 7408,
Bangladesh for allowing me the opportunity to continue my studies at the doctoral level. Special
thanks to Professor Dr. Md. Mortuza Ali, Professor Dr. Md. Rafiqul Islam Sheikh, and Professor
Dr. Md. Faruk Hossain at Rajshahi University of Engineering & Technology, Rajshahi,
Bangladesh who introduced me with plasma engineering, thin film technology, as well as research
opportunities in Japan. Special thanks to Md. Hasanuzzaman, Khulna University of Engineering
& Technology, Md. Abdul Majed Patwary, Comilla University, and Md. Mahbubur Rahman
Bhuiyan, Jahangirnagar University, Bangladesh.
Needless to mention, my beloved wife who has encouraged me always, whose love, support,
ii
patience, encouragement and sacrifice allowed me to finish my doctoral study. I am thankful to
my lovely son and daughter whose unconditional love to me and for his love, patience and sacrifice.
All of my heartfelt love to them. Finally, I deeply grateful to my parents, my father, is the person
who put the fundament my learning character, showing me the joy of intellectual pursuit ever since
I was a child; and my mother, is the one who sincerely raised me with her caring and gently love.
for their patience as well as moral support and endless love in long journey of my life. Their prayers,
encouragement, and advice have been and will always be a fortune for my life.
Finally, I would like to thank all of my laboratory friends for their supports as well as express
my gratitude to all Bangladeshi friends, Saga Moslem Society and international students in Saga
for making my Japan life more comfortable and enjoyable. This thesis is dedicated to my beloved
mother, father, wife, and my lovely son, ABMA Ishmam Amzad.
iii
Abstract
A high-density radio frequency (RF) magnetized sputtering plasma source with a rotational
square-shaped magnet arrangement for uniform target utilization has developed. Eight neodymium
rod magnets of 30 × 5 × 3 mm, where the connection between N-pole and S-pole magnets is one
side of the square, are mounted on a circular iron yoke disc and an iron cover of 5 × 3 × 1 mm is
also used for magnetic shielding of otiose magnetic fields from the permanent magnets. The
magnetic field simulation, the measurement of the target erosion and the time-averaged ion flux to
the target have been investigated for case (a) without iron cover, no air gap between N-pole and
S-pole magnets, case (b) with iron cover, no air gap, and case (c) with iron cover, 5 mm air gap,
respectively. It is found that the iron covers suppress the horizontal magnetic flux density and the
copper target utilization percentage increases from 74.15 % to 87.49 %. Moreover, by decreasing
the air gap between the shielded magnets, the copper target utilization percentage rises from
83.85 % to 87.49 %. The target utilization as well as the time-averaged ion flux to the target are
optimum for case (b).
A gyratory square-shaped capacitive radio-frequency (RF) discharge plasma sputtering
source is proposed for materials processing and functional film preparation, composed of magnet
arrangements consisting of eight neodymium bar magnets of dimensions 30 mm × 5 mm × 3 mm.
In order to evaluate its performance, two square-shaped magnetic arrangements were investigated:
case (a) without iron shielding and case (b) with iron shielding of dimensions 5 mm × 3 mm × 1
mm. The magnetic field simulation is analyzed, while the plasma discharge characteristics and the
film properties are measured. The film thickness and the resistivity profiles of case (b) are more
uniform than their corresponding profiles in case (a). The lowest electrical resistivity of the film
is 4.33 × 10-8 m at r = 30 mm for case (b), which is of the same order as the bulk resistivity of
the copper. The roughnesses of the film thickness profile for cases (a) and (b) are 24.4% and
7.2%, respectively. Using atomic force microscopy (AFM) analysis, the film surface for case (b)
was observed to show an improved smooth surface with reduced needle-shaped grain size, as well
as a lower surface roughness than that of case (a). The surface roughness of the films is
approximately 3.73 nm and 2.49 nm for case (a) and case (b), respectively. From the X-ray
iv
diffraction (XRD) patterns, the film texture, the relative intensity ratios of the (111) peak to the
(200) [I(111)/ I(200)] were found to be 13.76 and 4.08 for the cases (a) and (b), respectively.
To improve the target erosion near the edge, the outer ring-shaped RF magnetized plasma is
produced near the chamber wall by a monopole magnet scheme. Three monopole magnet schemes
such as the setups (a) R = 5 mm, (b) R = 20 mm and (c) R = 35 mm has been investigated are
chosen, where “R” is the gap distance between magnets in consecutive circles. Distributions of the
2D magnetic flux lines, absolute value of the horizontal magnetic flux density and discharge
voltage are investigated for the proposed setups to produce outer ring-shaped plasma. A high
luminous ring-shaped plasma is observed for (b) R = 5 mm, whereas multi-ring discharges are
observed for (b) R = 20 mm and (c) 35 mm. It is found that the electron temperature decreases
with increasing gas pressure for the all cases. The electron temperatures were 2.42, 1.71 and 1.15
eV at Ar gas pressure of 4 Pa for the setup (a), (b), and (c), respectively. The plasma density is
approximately same for the setups (b) and (c) at all gas pressure. The highest plasma densities
were 6.26×1015, 1.06×1016 and 1.11×1016 m-3 at 5 Pa for the setups (a), (b), and (c), respectively.
It is found that, the electron mean free path is 41.4, 63.17 and 84.66 mm at Ar gas pressure of 5
Pa for the setups (a), (b), and (c), respectively. Electron neutral collision frequency for case (a) R
= 5 mm is higher than that for case (b) R = 20 mm and case (c) R = 35 mm at a constant RF power
of 40 W and z = 13 mm axial distance from the target surface. Radial profile of ion saturation
current for case (b) R =20 mm is more uniform than that for case (a) R = 5 mm and case (c) R =
35 mm set up.
The capacitively coupled RF outer ring-shaped magnetized plasma discharge is developed
with a concentrically monopole arrangement of magnets to erode the target in a specific area, in
especial, near the chamber wall. The three concentric monopole magnet arrangements with a center
magnet, and magnets in setups (a) three circles, (b) two circles, and (c) one circle were investigated.
From the magnetic flux lines profiles, it was found that the magnetic flux density in component
parallel to the target surface has a peak magnitude in the outer circle of magnets for all setups.
Ring-shaped plasma in the specific outer area was observed. The ion saturation current, Iisat were
0.6 mA, 0.79 mA, and 0.46 mA, for setups (a), (b), and (c), respectively at r = 47 mm, where r =
0 mm is the center of the target. It was found that, Iisat is very high in the outer target region near
the chamber wall for setups (a) and (b), where Iisat for setup (c) decreases slowly. The results
showed that the target utilization could be controlled in the outer specific area near the wall.
v
A pulsed direct current (DC) discharge ring–shaped plasma source has been proposed using
single pole magnet arrangements, including a center magnet, with magnets in the setups (a) one
circle, (b) two circles, and (c) three circles. The 2D magnetic flux lines profiles, larmor–radii and
Hall parameters of the electrons and ions, electrical discharge characteristics, ion saturation current
profiles were investigated to characterize the proposed plasma. The electron larmor–radii, re were
0.17, 1.64, 5.82 mm for setups (a), (b), and (c), respectively. It was found that the highest electron
Hall parameters are approximately 561 at r = ±21 mm for setup (a), 544 at r = ±36 mm for setup
(b), and 297 at r = ±50 mm for setup (c). The strong ring–shaped plasma discharges was observed
for all setups. The typical discharge voltages were 1.0, 0.6, and 0.6 kV for setups (a), (b), and (c),
respectively. The ion saturation currents, Iisat were 1.44, 2.88, and 2.2 mA for setups (a), (b), and
(c), respectively at r = 45 mm and t = +10 μs. The Iisat of setup (b) is less fluctuating, whereas Iisat
of setup (c) is highly variable in all radial positions. Setup (b) has the best profile among the three
setups.
vi
Table of Contents
Acknowledgements i
Abstract iii
Table of contents vi
List of Symbols x
Chapter 1. Introduction 1
1.1. Plasma and Its Typical Features 1
1.1.1. Plasma Sheaths and Debye Length 4
1.1.2. Cyclotron Frequency and Larmor Radius 5
1.1.3. E × B Drift Motion 6
1.2. Thin Film Deposition by Plasma Sputtering 7
1.2.1. Principle of Plasma Sputtering 7
1.2.2. Merits and Demerits of Plasma Sputtering 9
1.3. Background Research on Plasma Sputtering 10
1.3.1. Synthesis of Functional Films by Plasma Sputtering 10
1.3.2. Previous Research for Improving Target Utilization and
their Problems
12
1.4. Problems of the Plasma Sputtering and Objectives of this Thesis 13
1.5. Structure of this Thesis 14
References 17
Chapter 2. Measurement Techniques and Methods 18
2.1. Major Experimental Apparatus 18
2.2. Ion Saturation Current, Plasma Density, and Electron
Temperature Measurement
19
2.3. Measurement of the Eroded Target Profile and Target Utilization 22
vii
Rate Calculation
2.4. Resistivity of the Prepared Film 24
2.5. Magnetic Field Analysis Simulation 26
2.6. Measurement of the Film Property 26
References 27
Chapter 3. Rotational Square-Shaped Arrangement of Rod Magnets for
Uniform Target Utilization
28
3.1. Introduction 28
3.2. Experimental Setup 29
3.3. Results and Discussions 33
3.3.1. Magnetic Field Profile Analysis of Square-Shaped
Schemes
33
3.3.2. Plasma Emissions and Discharge Characteristics 38
3.3.3. Profiles of the Target Erosion Depth and Ion Flux to the
Cu target
41
3.4. Conclusion 45
References 46
Chapter 4. Performance of a Gyratory Square-Shaped Capacitive Radio
Frequency Discharge Plasma Sputtering Source for Material Processing
48
4.1. Introduction 48
4.2. Experimental Setup Details 50
4.3. Results and Discussions 54
4.3.1. Magnetic Shielding Effect Square-Shaped Magnet
Arrangement
54
4.3.2. Plasma Discharge Voltage and Emission for Square-
Shaped Magnetized Plasma
57
4.3.3. Thickness and Resistivity Profiles of the Deposited Film
of the Gyratory Square-Shaped RF Magnetized Plasma
Sputtering Source
59
viii
4.3.4. Surface Morphology of the Film Deposited by the
Gyratory Square-Shaped RF Magnetized Plasma
Sputtering Source
61
4.3.5. Aluminum-Doped Zinc Oxide (AZO) Film Prepared by
RF Magnetized Plasma Sputtering Source with Square-
Shaped Rod Magnets
64
4.4. Conclusion 67
References 68
Chapter 5. Ring-Shaped Plasma for Target Utilization Obtained with Circular
Magnets Monopole Arrangement
70
5.1. Introduction 71
5.2. Experimental Methods 71
5.3. Results and Discussions 75
5.3.1. Magnetic Field Profile Analysis 76
5.3.2. Discharge Characteristics 79
5.3.3. Plasma Characteristics 82
5.3.4. Ion Saturation Current Profile 87
5.4. Conclusion 89
References 90
Chapter 6. Outer Ring-Shaped Radio Frequency Magnetized Plasma Source for
Target Utilization in Specific Area
92
6.1. Introduction 92
6.2. Experimental Setup Details 94
6.3. Results and Discussions 96
6.3.1. Profiles of Magnetic Field of the Monopole Setups 96
6.3.2. Emission of Outer Ring–Shaped Plasma 98
6.3.3. Outer Ring–Shaped Plasma Characteristics 100
6.3.4. Ion Saturation Current in Outer Ring–Shaped Plasma 103
6.4. Conclusion 105
ix
References 106
Chapter 7. Characteristics of a Ring–Shaped Pulsed DC Discharge Plasma
Source Using Single Pole Magnet Setups
108
7.1. Introduction 108
7.2. Experimental Arrangements 110
7.3. Results and Discussions 113
7.3.1. Simulation of Single Pole Magnet Setups for Ring-
Shaped Pulsed Discharge
113
7.3.2. Electrical Characteristics of the Ring–shaped Pulsed
Discharge Plasma
121
7.3.3. Discharge Characteristics of the Ring–shaped Pulsed
Discharge Plasma
123
7.3.4. Ion Saturation Currents Profiles 126
7.4. Conclusion 129
References 131
Chapter 8. Summary and Suggestions for Future Plan 134
8.1. Summary and Conclusions 134
8.2. Suggestions for Future Works 137
List of Refereed publication papers and conferences 138
x
List of Symbols
Parameter Definition Unit
a Maximum of target erosion height µm
A Probe area m2
Af Cross section of the deposited film resistance m2
B Radial interval between consecutives target erosion depth m
B Magnetic flux density G
Bx Horizontal magnetic flux density to target G
Bz Vertical magnetic flux density to target G
E Electronic charge C
H Hall parameter Unit less
he Electron hall parameter Unit less
hi Ion hall parameter Unit less
Ie Electron current detected by the probe mA
Iesat Electron saturation current detected by the probe mA
Iisat Ion saturation current detected by the probe mA
Ip Probe current mA
K Boltzmann’s constant J/K
m Electron mass kg
M Ion mass kg
n Charge number density m-3
nAr Argon gas number density m-3
nCu Atomic number density of the copper target m-3, cm-3
ne Electron number density m-3, cm-3
ni Ion number density m-3, cm-3
q Electric charge C
xi
r Radial position from center of the target mm
R Gap distance between magnets in consecutive circles mm
Ri Ammeter internal resistance Ω
Rv Voltmeter internal resistance Ω
Rf Deposited film resistance Ω
rL Larmor radius m
rLe Electron larmor radius m
rLi Ion larmor radius m
Rsp Sputtering rate µm/hour
Si Area occupied by erosion at target m2
Sp Area of the probe used to detect the ion saturation current m2
Te Electron temperature eV
TU Target utilization percentage %
V Charge velocity component m/s
ve Electron velocity component m/s
VB Biasing voltage V
VRF RF discharge voltage (peak - to - peak value) Vpp
VSB Self-biased voltage V
λD Debye length m
ωc Cyclotron frequency Hz
ϵo Permittivity of free space F/m
γsp Yield of sputtered atoms per incident ion Unit less
νm Charged particle neutral collision frequency Hz
νme Electron neutral collision frequency Hz
νmi Ion neutral collision frequency Hz
1
Chapter 1. Introduction
1.1. Plasma and Its Typical Features
A plasma is a collection of free charged particles (electrons and ions) moving in a
random direction that are electrically neutral (ne ≈ ni) where ne and ni are the electron and ion
densities, respectively [1-2]. Artificially plasma can be generated by heating or be subjecting a
neutral gas to a strong electromagnetic field to the point an ionized gaseous particles become
electrically conductive, and the behavior of the matter is dominated by the electromagnetic
fields in long-range [3]. Plasma is matter heated beyond its gaseous state, heated to a
temperature so high that atoms are stripped of at least one electron in their outer shells so that
what remains are positive ions in a sea of free electrons. That is, as the temperature increases,
the atoms become more energetic and transform matter in the sequence: solid, liquid, gas, and
finally plasma which justifies the title, "fourth state of matter".
Fluorescent lights contain plasmas and Lightning is an example of plasma present at
Earth's surface. Aside from those, we do not often encounter plasmas in everyday life. Outside
the earth in the ionosphere or outer space, however, almost everything is in the plasma state.
Some examples are aurora borealis, solar winds, magnetospheres of earth and Jupiter, gaseous
nebulae, galactic arms, quasars, pulsars, novas, black holes, fluorescent lamp, fires and TVs
[2]. The plasma ion engine is used for the flight of the space probe "Hayabusa". On the other
hand, plasma processing is the most widely used chemical process in microelectronic industry
for thin film deposition and etching. Its application expands to surface modification, flat panel
display fabrication, solar cell, plasma spray and plasma micro-discharge and many other
growing areas [4]. The studies such as sterilization and sterilization, the growth stimulation of
the plant are pushed forward recently. It is highly advanced the imminent product of all of you
by this plasma process technology.
Figure 1.1 shows the schematic view of a plasma showing charged particles movements
in the random directions. In general, plasmas have the following features [1]: (1) they are driven
by voltage or current source, electrically; (2) charged particles collisions with neutral gas
molecules are significant; (3) there are boundaries at which surface losses are significant; (4)
in the steady state ionization of neutrals sustains the plasma; and (5) the electrons and the ions
are not in thermal equilibrium.
2
Figure 1.1. Schematic view of a plasma showing charged particles movements in the andom
directions.
Figure 1.2 shows a typical radio-frequency (RF) plasma discharge system. It involves
a voltage source that flows current through a low-pressure gas between two parallel conducting
plates. The gas break down to form a typical plasma, usually weakly ionized. Plasma is formed
between the electrodes. As shown in the Fig. 1.2, the process chamber can be surronded by dc
multipole magnetic fields to improve plasma confinement near the chamber surfaces, while
providing a magnetic near-field-free plasma atmosphere at the wafer. Such arrangements are
often called remote sources [1]. Sometimes, the source and process chambers are more complex.
For example, the wafer is kept very near to the source exit, to obtain increased ion and radical
fluxes, reduced spread in ion energy, and improved process uniformity. But the wafer is then
exposed to higher levels of damaging radiation. Although the need for low pressures, high
fluxes, and controllable ion energies has motivated high-density source development, there are
many issues that need to be resolved [1]. A critical issue is achieving the required process
uniformity over 300 mm wafer diameters. In contrast to the nearly one-dimensional geometry
of typical rf diodes (two closely spaced parallel electrodes), high-density cylindrical sources
can have length-to-diameter ratios of order or exceeding unity. Plasma formation and transport
in such geometries are inherently radially nonuniform [1].
3
Figure 1.2. A typical radio-frequency plasma discharge system.
There is a wide range of densities and temperatures for both artificially processed and
space plasmas. Low-pressure plasma discharges characterized by plasma density, n ≈ 1014 -
1019 m-3, and electron temperature, Te ≈ 1 – 10 eV, which are used as chemically reactive
etchants, functional film depositions precursors. In general, while energy is delivered in the
form of bombarding ions to substrate surface, the energy flux enhances the chemistry at the
substrate surface, do not heat the substrate. The functional gas pressure for these discharges are
≤ 1 Pa 133.3 Pa and subjected of this thesis.
The high-pressure discharges have Te ≈ 0.1 – 2 eV and n ≈ 1020 – 1025 m-3 are also used
in plasma processing. The light and heavy particles are more nearly in thermal equilibrium
satisfying the condition Ti ≤ Te. These discharges are used mostly to distribute heat to the
substrate surface to increase surface reaction rates, to melt, evaporate materials, to weld
refractory materials. The operating pressures are approximately atmospheric pressure,
1.01325×105 Pa. Shock tubes, focus, high-pressure arcs and laser plasma are the examples of
high-pressure discharges have plasma density approximately n ≈ 1020 m-3. Fusion reactor
plasmas (n ≈ 1020 – 1021m-3, and Te ≈ 3.5 – 4 eV), alkali metal plasmas (n ≈ 1015 – 1018 m-3,
and Te ≈ 1 eV), flames (n ≈ 1014 m-3, and Te ≈ 1 eV), earth ionosphere (n ≈ 1010 1014 m-3, and
Te ≈ 2 eV) and laser plasma (n ≈ 1025 – 1027m-3, and Te ≈ 2 – 3 eV), are typical examples of
the plasma discharges.
4
1.1.1. Plasma Sheaths and Debye Length Figure 1.3 shows the profile of the plasma densities after formation of the two sheaths
near the chamber wall. The condition ne ≈ ni is called as quasi-neutrality and the most important
characteristics of a plasma [1-2]. Plasmas are quasi-neutral (ne ≈ ni), and are joined to chamber
wall surfaces across thin positively charged layers called sheaths. The plasma potential varies
slowly in the plasma but rapidly in the sheath. The plasma density is relatively flat in the center
and falls sharply near the sheath edge. The electron densities fall rapidly at chamber wall
surface. If the sheaths potential drops are unequal, the electron fluxes will be unequal. The
Bohm criterion states that ions must steam in the sheath with a velocity at least / / [2].
Figure 1.3. Profile of the plasma densities after formation of the two sheaths near the
chamber wall.
A fundamental characteristic of the behaviors of a plasma is its ability to shield out
electric potentials of individual charged particles that are applied to it. Debye length indicates
that the distance scale over which significant charge densities can spontaneously exist. The
5
characteristic length scale in a plasma is the Debye length can be expressed by the following
equation [1-2].
⁄
(1.1)
It is assumed that plasma having equilibrium densities ne = ni = n0. In practical units, it can be
written as following form [1]:
743 ⁄ (1.2)
With Te in electron volts and ne in cm-3. For example, Debye length is 0.14 mm for Te = 4 eV
and ne = 1010 cm-3.
1.1.2. Cyclotron Frequency and Larmor Radius
If the plasma is inserted in a DC magnetic field (B-field) the motions of the charged
particles are affected by the B-field and makes the plasma an anisotropic medium, with a
preferred direction along B. That is, the charged particle, electron rotate along the B-field as
shown in figure 1.5. As long as the ion or electron of charge q is moving, it experiences a
Lorentz force qv×B, which is perpendicular to both velocity and the field. This force has no
effect on the velocity component to B, but in the perpendicular plane, it powers the particle to
rotate in a cyclotron orbit. The frequency of this circular motion, the cyclotron frequency is
represented by the following equation [2].
⁄ (1.3)
The radius of the circle of rotation called Larmor radius or gyro-radius . The cyclotron
frequency is independent of velocity, however, gyro-radius depend on velocity. If is the
velocity component in the plane perpendicular to B, a particle completed an orbit a length 2
in a time 2 ⁄ , so , alternatively can be written by the following formulae [2].
6
⁄ (1.4)
⁄ (1.5)
That is, the presence of B-field controls charged particles, in especial, electrons so that
electrons are magnetized. The magnetized electrons play an important role to produce a high-
density plasma that can enhance the deposition rate for plasma processing such as plasma
sputtering. Figure 1.4. shows the rotation of electron in the presence of the magnetic field. The
Larmor radius can be controlled by the velocity, v and the magnetic field, B, respectively.
Figure 1.4. Rotation of electron in the presence of the magnetic field.
1.1.3. E × B Drift Motion
The direction of electric field E and the behavior of magnetic flux density B plays an
important role in performing the magnetized discharge based on E×B drift motion, where E
and B are the electric field perpendicular to the target and magnetic flux density parallel to the
target, respectively. The direction of E × B drift motion in the presence of electric and magnetic
field is shown in Fig. 1.5. Moreover, it can be demonstrated by the right hand rule. If the
magnetic fields are so strong that both ions and electrons have Larmor radii much smaller that
the plasma radius. So that the particles guiding centers drift across B-field, in response to
applied electric fields E, the perpendicular component to B-field. The drift velocity can be
7
found using the following Eq. (1.6). The velocity component parallel to B-field is unaffected
by E-field. Moreover, the drift velocity, v is perpendicular to both E-field and B-field and have
the same value for ions and electrons.
| |
(1.6)
Figure 1.5. Direction of E × B drift motion in the presence of electric and magnetic field.
1.2. Thin Film Deposition by Plasma Sputtering 1.2.1. Principle of Plasma Sputtering
Sputter deposition is a physical vapor deposition (PVD) method of thin film deposition.
This process consists of ejecting material from a target surface onto a substrate. Sputtered
atoms ejected from the target surface have a wide energy distribution, typically up to tens of
eV (100,000 K). The number of atoms ejected or “Sputtered off” from the target surface is
called the sputter yield. The sputter yield depends on the energy and incident of angle of the
bombarding ions, the relative masses of the ions and target atoms, and the surface binding
energy of the target atoms. Plasma processing such as plasma CVD and sputtering are widely
used for fabrication of microelectronic thin film preparation. Magnetron sputtering has become
the process of choice for the deposition of a wide range of industrially important coatings.
Figure 1.6 shows the principle of magnetron sputtering demonstrating the direction of electric,
magnetic field, charged particle motions and target erosion. Magnetron Sputtering is a Plasma
8
Vapor Deposition (PVD) process in which a plasma is created and positively charged ions from
the plasma are accelerated by an electrical field superimposed on the negatively charged
electrode or "target". The positive ions are accelerated by potentials ranging from a few
hundred to a few thousand electron volts and strike the negative electrode with sufficient force
to dislodge and eject atoms from the target. These atoms will be ejected in a typical line-of-
sight cosine distribution from the face of the target and will condense on surfaces that are
placed in proximity to the magnetron sputtering cathode. A magnetic field is applied at right
angles to the electric field by placing large magnets behind the target. This traps electrons near
the target surface and causes them to move in a spiral motion until they collide with an Ar atom. To increase deposition rates, magnets are used to increase the percentage of electrons that take
part in ionization events, increasing the ionization efficiency. Orbital motion of electrons
increases probability that they will collide with neutral species and create ions.
9
Figure 1.6. Principle of magnetron sputtering demonstrating the direction of electric,
magnetic field, charged particle motions and target erosion.
1.2.2. Merits and Demerits of Plasma Sputtering
Magnetron Sputtering Sources are very familiar plasma source for functional thin film
preparation such as metal, oxide, nitride and nano-materials thin films. The merits are high-
speed deposition rate, high-quality thin film property. However, the major demerits are non-
uniform target erosion profile, lower target utilization rate such as 30 - 40% due to localized
ring-shaped high-density plasma. So that, it is needed to develop a novel plasma sputtering
source with improved target utilization percentage, erosion rate and uniform film property.
10
1.3. Background Research on Plasma Sputtering
1.3.1. Synthesis of Functional Films by Plasma Sputtering
Cu films of 290–350 nm were prepared by DC magnetron sputtering, the target voltage
of 400 V, and Ar gas pressure was introduced at 0.5, 1.0, and 1.5 Pa, respectively by H. Qiu et
al [5]. It was found that the target current increased from 69 to 200 mA with increasing Ar
pressure, and the target voltage was settled at 400 V. The amount of larger grains decreases
with an increase in Ar pressure while the resistivity of the films increases with increasing Ar
pressure. The roughness of the film surface was approximately 5 nm, which was independent
of Ar pressure [5]. The crystalline orientation of the prepared Cu film shifted slightly from the
[1 1 1] direction to the [2 2 0] direction with increasing Ar pressure [5].
Cu doped ZnO films at various doping concentrations of Cu (0, 5.1, 6.2 and 7.5%) by
simultaneous RF and DC magnetron were prepared by A. Sreedhar et al [6]. The structural,
electrical and optical properties of ZnO films was discussed in detail. X-ray diffraction (XRD)
patterns show that the films were polycrystalline in nature wurtzite structure toward c-axis and
Atomic force microscopy (AFM) results indicate that the films displayed needle-like shaped
grains throughout the substrate surface. The electrical resistivities were found to be increased
with an increase of copper content from 0 to 7.5% [6]. Films showed an average optical
transmittance about 80% in the visible region and optical band gap values decreased from 3.2
to 3.01 eV with increasing of Cu doping content from 0 to 7.5%, respectively [6]. However,
the optical band gap values were decreased from 3.2 to 3.01 eV with increasing the copper
doping concentration from 0 to 7.5% in the ZnO host lattice respectively [6].
The size effect on the resistivity of evaporated copper films, ranging in thickness from
9 nm to 167 nm, was determined experimentally from the sheet resistance and the physical
thickness [7]. In combination, to determine the mean grain size of the grains in the plane of the
film, the Electron Back Scatter Diffraction (EBSD) and the XRD methods were used, A. E.
Yarimbiyik et al proposed Matthiessen’s rule can be used to measure the thickness of a copper
film and [7]. The resistivity of these films increased with decreasing film thickness [7].
Al-doped ZnO (AZO) thin films have been prepared by radio frequency (RF)
magnetron sputtering and applied RF power was varied in the range 600–1200 W by A.
Spadoni et al [8]. The effects of RF power on structural, electrical and optical properties were
examined by XRD analysis, Hall measurements and UV–vis–NIR spectrophotometry
respectively. It was found that increasing the RF power, AZO films having a preferential
growth orientation along (002) direction, showed a decrease of the lattice distance indicating a
11
less defected structure [8]. It was observed that the increase of the RF power made a higher
optical absorption by free carriers, coupled with an increase of the band gap value. Resistivity
varied from 1.1 × 10−3 Ω cm at 600 W down to a minimum value of 5.6 × 10−4 Ω cm at 1200
W, whereas the carrier density increased up to 1 × 1021 cm−3. Lattice defect variation of AZO
films was analyzed by photoluminescence (PL) measurements. It was hypothesized that at
higher RF power a more effective diffusion phenomenon can give more effective Al doping
and less amount of zinc vacancies. As a consequence, Al atoms are more effectively trapped
into the structure [8].
S. Rahmane et al deposited Al-doped ZnO (AZO) films on glass and silicon substrates
by RF magnetron sputtering technique at room temperature [9]. The effects of thickness on the
structural, optical and electrical properties were investigated. It was found that the electrical
resistivity decreases with the increase of the film thickness and the smallest measured value
was 8 ×10-4 Ω cm for the 1500 nm thick film. All the deposited films showed a crystalline
wurtzite structure with a strong preferred (002) orientation. It was observed that the intrinsic
compressive stress decreased with the increase of the film thickness [9]. The obtained AZO
films had an average transmittance greater than 90% in the visible region and films have an
optical band gap between 3.32 and 3.49 eV depending on the film thickness [9].
Cu film has been proposed because of its low resistivity, high chemical stability, and
excellent electromigration resistance B. H. Wu et al [10]. The films deposited on a silicon
substrate by varying the substrate bias voltage using high power pulsed magnetron sputtering
(HPPMS). It was found that the substrate bias increased from - 17.3 V to - 100 V, the electron
current decreases slowly, while the ion current increases, and steadying at - 50 V [10]. It was
observed that the Cu films prepared at - 50 V, and -100 V exhibited a higher tensile stress and
a superior (111) texture. The electrical resistivity of the deposited Cu films was found to be a
minimum value of 1.79 μΩcm at - 100 V [10].
A. S. Christiansen et al investigated the nitrogen dissociation and plasma parameters
during radio frequency sputtering of lithium phosphorus oxynitride thin films in nitrogen gas
by mass appearance spectrometry, electrostatic probes, and optical emission spectroscopy [11].
Despite lower plasma density, the film grows quicker at a lower pressure where the higher
plasma potential, translated into higher energy for imposing ions on the substrate which
resulted in a compact and smooth film structure. It was found that low pressure (5 mTorr) and
moderate power (100 W) were most beneficial for the growth of good quality films with a high
ionic conductivity [11]. Increasing the RF power (300W) resulted in a poor film quality due to
cracks and dislocation of large clusters from the target material, while increasing the nitrogen
12
pressure (50 mTorr) resulted in a lower deposition rate and a rough microstructure with
volcano-shaped structures on the surface. Higher pressures showed much less nitrogen
dissociation and lower ion energy with thinner films, less ionic conductivity and poor film
structure with large roughness [11].
A highly transparent field emitter was achieved by Ar+ ion irradiation onto highly
transparent and conducting ZnO films deposited on glass substrates by Zurita Zulkifli1 et al
[12]. The deposited flat ZnO films before ion irradiation, which showed 90% transmittance and
186 Ω/cm sheet resistance, showed no field emission current up to 15 μVm-1 [12]. The
nanocone size was less than the wavelength of visible light and the transmittance was
maintained at 86% for the ion-irradiated ZnO film [12]. It was observed that the field emission
properties of the transparent nanoconed ZnO were promising compared to other nontransparent
nanostructured ZnO [12].
An inverted gapped-target magnetron sputtering device for deposition of ferromagnetic
thin films under energetic conditions has been developed by P. Poolcharuansin et al [13]. The
side view and the front view pictures during the inverted plasma discharge with a power of 80
W and with a pressure of 2 Pa were discussed [13]. No cracking of the film surfaces was found
in the sputtering of nickel films without substrate heating or biasing. Langmuir probe and ion
energy measurements confirmed that plasma ions with a density of around 1016 m-3 and the
almost energies of approximately 200 eV can be reached at the substrate [13].
1.3.2. Previous Research for Improving Target Utilization and their Problems
A wider eroded and higher target utilized magnetron sputtering system for
ferromagnetic nickel (Ni) target using a large, tall and eccentrically rotating tilted center
magnet was proposed by T. Iseki et al [14]. The target utilization was found approximately
49% for 5-inch and 4-mm thick Ni target, which is 8% better than when a non-tilted center
magnets were used. It was observed that the rotating mechanism of a yoke magnet could be
separated from the cooling water and the target could be cooled more effectively and the center
non-eroded area was decreased [14].
T. Iseki et al investigated the dependence of the magnetic flux density, erosion
uniformity, and target utilization on the yoke magnet tilt angle in a planar magnetron sputtering
system, using a rotating, tilted, unbalanced, asymmetrical magnet [15]. The magnetic flux
density distributions were measured two-dimensionally on the target surface. As the yoke
magnet tilt angle increased from 0 to 8 degree, utilization of a 5-inch target linearly increased
from 60 to 80%. On the other hand, with an elliptical outer yoke, the target utilization was
13
approximately 70%, regardless of the yoke magnet tilt angle [15]. It was found that the
deposition rate when using the elliptical outer yoke was 1.2 times faster than that of when using
the circular outer yoke at the same magnet tilt angle [15]. However, to make a tilt angle in a
yoke surface, an unbalanced magnet is arranged on the magnet holder surface. The number of
magnets are different from one side to another side. These effect make short system life.
Y. Ohtsu et al proposed the racetrack-shaped RF magnetron plasma with weak rubber
magnets (ferrite and neodymium) for the full utilization of the circular target and the reduction
of the magnet weight [16]. The magnetic field simulations for the ferrite rubber magnets, the
neodymium rubber magnets, and the neodymium rubber magnets including the neodymium
metal magnets were investigated [16]. The radial profile of the erosion depth was roughly
constant for r = 20 mm and then decreases slowly away from the center for an RF power of 40
W, Ar gas pressure of 2 Pa and a sputtering time of 4 h. It was found that the target utilization
was approximately 72% estimated from the target erosion profile [16].
1.4. Problems of the Plasma Sputtering and Objectives of this Thesis
Magnetron Sputtering Sources are very familiar plasma source for functional thin film
preparation such as metal, oxide, nitride and nanomaterials thin films. Recently, RF magnetized
plasma sources have been widely used in microelectronics, such as magnetic films, surface
treatment and cleaning, diamond-like carbon, biomaterial thin films, flat panel display
fabrication, transparent conductive oxide film preparation for solar cells and mobile phones,
and many other rapidly growing areas. In particular, RF magnetron plasma sources have
become an attractive tool for functional film preparation. In the conventional magnetron system,
the target material is not effectively used because high-density plasma is localized on the target
surface. The target utilization is very low approximately at 20-30% [5].
From the practical viewpoint of the limited resources, the utilization of the target
material is necessary. The symmetrical magnets magnetron sputtering method with one inner
magnet and two outer annular magnets facing each other were investigated. The maximum
target erosion rate was 57% [17]. The rotating magnet sputtering has also been proposed by
rotating helical magnets for increasing the target utilization efficiency. Using a rotating
unbalanced and asymmetrical magnet, a flat erosion-sputtering method has been developed.
The estimated target utilization had a value of 80% and 77% for five, four-inch aluminum target
material, respectively [13-14]. The target utilization efficiency was increased from 73.6% to
86.3% when iron pole pieces were used in the rotating cruciform arrangement of neodymium
magnets [17-18].
14
In general, near the chamber wall, the plasma potential and the plasma density is very
small because an ion sheath exists near the wall [1, 19]. The plasma potential varies slowly in
the plasma, but rapidly in the sheath region. Only in the sheath region, the quasineutrality
property cannot be satisfied. The plasma density profile is relatively flat in the center and falls
sharply near the sheath edge. It is required to produce (1) outer ring-shaped and (2) specific
area plasmas for obtaining (1) uniform, (2) high-density plasma as well as (3) convenient outer
target area erosion profile near the chamber wall. Moreover, plasma processing has various
problems on plasma equipment, thin film preparation and so on. In practical industrial
application, the outer width of the target is large.
Therefore, it is required high-density plasma discharge in a specific area and also in the
outer region of the chamber to obtain convenient outer target area erosion profile near the
chamber wall. Moreover, to deposit a functional thin film in a specific area and near the
chamber wall, the target utilization in a specific area is required from the viewpoint of target
utilization. However, the conventional magnetron plasma has an issue that the target erosion is
not uniform owing to high-density localized plasmas.
In this thesis, the developments of novel plasma source equipment with uniform target
utilization are investigated for the growing plasma processing applications. The following aims
and objectives of the research are listed below.
Finding the way of increasing target utilization percentage
Making a uniform target erosion over the entire area of the target
Depositing a uniform functional film property
Producing a specific area plasma for target utilization near the chamber wall
Obtaining a pulsed dc discharge plasma for target utilization
1.5. Structure of this Thesis
The structure of the thesis is organized into separate chapters. Chapter 2 discusses the
experimental methods and measurement techniques used during the doctoral study as well as
simulation and analysis methods used to obtain and interpret information. Various kinds of
measurement tools and diagnostics are presented.
Chapter 3 describes a high-density radio frequency (RF) magnetized sputtering plasma
source with a rotational square-shaped magnet arrangement for uniform target utilization. The
experimental setup for the proposed RF magnetized plasma sputtering system is explained in
detail. The magnetic field profiles obtained based on the squared-shaped magnets, the plasma
15
emission and discharge characteristics, the copper target erosion profile and ion saturation
currents are investigated explicitly. Finally, the results obtained in this chapter are summarized.
Chapter 4 is dedicated to investigating the performance of a gyratory square-shaped RF
magnetized plasma sputtering source for materials processing and different functional film
preparations. A radial profile and microstructure of the deposited films are investigated under
two square-shaped magnet schemes. The experimental setup, the results and discussions on the
effects of the magnetic shielding material in a square-shaped magnet arrangement on the
discharge characteristics, plasma emission luminescence, thickness and resistivity profile, as
well as the surface morphology of the film deposited by the gyratory square-shaped RF
magnetized plasma sputtering source is explained in detail. Finally, the results obtained in this
chapter are summarized.
Chapter 5 is focused to make an RF magnetized outer ring-shaped plasma sputtering
source with a concentrically monopole arrangement of magnets at various different magnet-
gap distances for the specific area target utilization. The experimental setup, the effects of
monopole schemes, the discharge characteristics, electron temperature, plasma density and
electron collision frequency in the monopole arrangements are explained in detail. Finally, the
results obtained in this chapter are summarized.
Chapter 6 is devoted to producing outer ring-shaped and specific area plasmas for
obtaining high-density plasma and outer target erosion profile near the chamber wall. The
proposed experimental setup, the results and discussions on the effects of the three setups are
investigated in a monopole circular magnet arrangement of the magnetic flux lines and their
profiles, discharge characteristics, plasma discharge luminescence, plasma density, and ion
saturation currents are explained in detail. To conclude, the results obtained in this chapter are
summarized. Chapter 7 is dedicated to characterizing a ring–shaped pulsed DC discharge plasma
using single pole magnet setups in a specific area for obtaining outer target erosion profile near
the chamber wall. The proposed experimental setup of pulsed DC discharged plasma, the
results and discussions on the effects of the three setups are investigated in a single pole magnet
arrangement, 2D magnetic flux lines and their profiles, gyro–radius of the electrons and ions,
Hall parameters of the electrons and ions, discharge characteristics, ion saturation current
profiles are explained in detail. As a final point, the results obtained in this work are
summarized.
16
Chapter 8 is summarized based on chapter 1 to chapter 7. Few possible suggestions for
future works are recommended at the end of this thesis. List of refereed publication papers and
contributed conferences and symposiums are listed.
17
References
[1] M. A. Lieberman, and A. J. Lichtenberg, “Principles of Plasma Discharges and Material
Processing”, 2nd ed. John Wiley & Sons, Inc., New York, 2005.
[2] F. F. Chen, and J. P. Chang, “Lecture Notes on Principles of Plasma Processing”, Plenum,
New York, 2002.
[3] A. I. Morozov, “Introduction to Plasma Dynamics”, CRC Press, 2012.
[4] T. Makabe “Advances in Low Temperature RF Plasmas, Basis for Process Design, Elsevier,
2002.
[5] H. Qiu, F. Wanga, P. Wua, L. Pana, and Y. Tiana, Vacuum, 66, 447–452, 2002.
[6] A. Sreedhar, J. H. Kwon, J. Yi, J. S. Kim, and J. S. Gwag, Materials Science in
Semiconductor Processing, 49, 8–14, 2016.
[7] A. E. Yarimbiyik, H. A. Schafft, R. A. Allen, M. D. Vaudin, and M. E. Zaghloul,
Microelectronics Reliability, 49, 127–134, 2009.
[8] A. Spadoni, and M.L. Addonizio, Thin Solid Films, 589, 514–520, 2015.
[9] S. Rahmane, M. S. Aida, M. A. Djouadi, and N. Barreau, Superlattices and Microstructures,
79, 148–155, 2015.
[10] B. H. Wu, J. Wu, F. Jiang, D. L. Ma, C. Z. Chen, H. Sun, Y. X. Leng, and N. Huang,
Vacuum, 135, 93 – 100, 2017.
[11] A. S. Christiansen, E. Stamate, K. Thyden, R. Younesi, and Peter Holtappels, Journal of
Power Sources, 273, 863 – 872, 2015.
[12] Z. Zulkifli1, S. Munisamy, M. Z. M. Yusop, G. Kalita, and M. Tanemura, Japanese Journal
of Applied Physics, 52, 11NJ07, 2013.
[13] P. Poolcharuansin, P. Laokul, N. Pasaja, A. Chingsungnoen, M. Horprathum, P.
Chindaudom, and J. W. Bradley, Vacuum, 141, 41-48, 2017.
[14] T. Iseki, H. Maeda, T. Itoh, Vacuum, 82, 1162–1167, 2008.
[15] T. Iseki, Vacuum, vol. 84, pp. 339–347, 2010.
[16] Y. Ohtsu, S. Tsuruta, T. Tabaru, and M. Akiyama, Surface Coatings Technology, 307,
1134–1138, 2016.
[17] Y. Ohtsu, M. Shigyo, M. Akiyama, T. Tabaru, Vacuum, 101, 403 – 407, 2014.
[18] T. Ide, M. A. Hossain, Y. Nakamura, and Y. Ohtsu, Journal of Vacuum Sci. Technol. A:
Vacuum, Surfaces, and Films, 35, 061312, 2017.
[19] P. J. Kelly, R. D. Arnell, Vacuum, 56, 159 – 172, 2000.
18
Chapter 2. Measurement Techniques and Methods
2.1. Major Experimental Apparatus
Major apparatus used to do the experiment in this research showing matching network,
vacuum system, external motor circuits, oscilloscope, and plasma discharge chamber are
shown in Fig. 2.1. The experiments were performed in a stainless-steel cylindrical vacuum
chamber with an outer diameter of 235 mm, an inner diameter of 160 mm and a height of 195
mm. The discharge chamber was evacuated to a base pressure of 10-5 Pa by a turbo molecular
and an oil rotary pump [1-3]. Argon (Ar) gas was used as sputtering gas by regulating a flow
meter. An RF power source at 13.56 MHz was applied to the target via an impedance matching
network. A copper plate of 3 mm thickness was used as the target to measure the sputtering
characteristics. An aluminum disc was only used to measure plasma parameters in order to
avoid thin film deposition on the Langmuir probe. The discharge voltage between the RF
powered electrode and the grounded vacuum wall was measured by a high-voltage probe
connected with oscilloscope.
The iron yoke is set up with a gap of 1 mm, and 0 mm from the copper target plate for
rotational case and stationary case, respectively. The circular iron yoke disk with the magnet
arrangement was rotated by an iron yoke shaft by an external motor. A precision surface tool
was used to measure the erosion depth of the copper target as a basis to calculate the target
utilization percentage. The plasma parameters such as the time-averaged ion saturation currents
are measured to obtain the ion flux to the target by a tiny tungsten wire probe of 1.0 mm in
diameter and 10 mm in length. In order to avoid the influence of the RF potential oscillations
on the probe current-voltage characteristics, the probe wire was connected to an LC filter circuit
[1-2]. The ion flux to the target is calculated from the ion saturation current by negatively
biasing the probe including a resistance of 100 Ω.
19
Figure 2.1. Major apparatus used to do the experiment in this research showing matching
network, vacuum system, external motor circuits, oscilloscope, and discharge chamber.
2.2. Ion Saturation Current, Plasma Density and Electron Temperature
Figure 2.2 shows ion saturation current and current-voltage characteristics curve
measuring equipment. Langmuir probe and DC power supply were used. The biasing voltage
was changed from -70 V to +70 V. The Langmuir probe was compensated with an LC parallel
filter circuit to avoid the influence of the RF plasma potential fluctuations on the probe current-
voltage characteristics. The probe measurements were carefully performed so as to minimize
the disturbance to the stable plasma [1-3].
20
Figure 2.2. Ion saturation current and current-voltage characteristics curve measuring
equipment.
Figure 2.3 shows a practical current-voltage characteristics curve showing its various
parts. The point at which the curve crosses the V axis is called the floating potential Vf. At any
potential, where electron saturation currents start is defined as Vp. The electron temperatures,
the plasma density are estimated using current-voltage characteristics curve. According to the
probe theory [1-2], the ion saturation current is proportional to the ion flux to the target.
Moreover, the plasma density can be estimated from ion saturation, which can be expressed as
the following equation.
0.6 / (2.1)
21
Figure 2.3. A practical current-voltage characteristics curve showing its various parts.
At floating potential, electrons and ions currents are equal that is Ie = Ii. So that total
current is zero. The electron current can be expressed as Eq. (2.2). Taking logarithm and
making simplification Eq. (2.2) can be written in the following form in Eq. (2.3) and (2.4). So
that the inverse slope of the logarithmic electron probe current with respect to biasing voltage,
VB (in volts) gives electron temperature, Te directly.
exp (2.2)
ln (2.3)
ln
(2.4)
The measurement of ion saturation current, Iisat is the simplest and best way to
determine the plasma density, n. The sheath around a negatively biased probe is so thin that the
area of the sheath edge is essentially the same as the area of the probe tip itself at higher plasma
densities more than 1016 m-3. Eq. (2.1) indicates a constant ion saturation current, which can
occur only for flat probes in which the sheath area cannot increase as the probe is functioned
22
more and more negative. In practice, ion saturation current, Iisat usually has a slope to it because
the Iisat has to come from a disturbed volume of plasma (the presheath). Plasma density can be
calculated from the following Eq. (2.5).
0.6 /
(2.5)
2.3. Measurement of the Eroded Target and Target Utilization Rate Calculation
A precision surface profile tool (Mitutoyo SJ-400) was used to measure the radial
profile of the thickness of the deposited Cu film. Figure 2.4 shows the target erosion depth and
film thickness measuring apparatuses showing eroded target and moving step profiler. For
erosion depth measurement, measurements are carried out using a step meter (Mitutoyo SJ-
400). A precision tape was attached to a part of the target material. The target material was set
for the experiment for plasma sputtering. After the experiment, the tape was taken off. The
mechanical step at the tape boundary between tape part and the sputtered part to measure the
difference. Scanning was done in the direction perpendicular to the target surface. Scan in the
direction of the erosion surface from the part where it is eroded. The thin film thickness is
obtained from the output data obtained by the stylus profilometer. In fact, it captures the fine
unevenness of the material surface with high definition.
The target utilization percentage was estimated based on the ratio of the target erosion
volume to the volume calculated under the assumption that the target is completely eroded at a
maximum erosion depth [4]. Target erosion depth profile and calculation of the target
utilization percentage is shown in Fig. 2.5. The target utilization percentage is calculated using
following Eq. (2.6).
2 / 100 % (2.6)
23
Figure 2.4. Target erosion depth and film thickness measuring apparatuses showing eroded
target and moving step profiler.
24
Figure 2.5. Target erosion depth profile and calculation of the target utilization percentage.
2.4. Resistivity of the Prepared Film
The resistivity of the deposited films was measured by a standard four terminals probe
method at a room temperature. Separate pairs of current-carrying and voltage-sensing
electrodes were used to make more accurate measurements by four-terminals probe method.
Figure 2.6 shows Four terminals probe method to measure deposited film resistivity mentioning
(a) schematic diagram and (b) electrical circuit diagram. The deposited film resistance can be
calculated using current voltage characteristics curve. However, the film resistance is expressed
as following Eq. (2.7), where Af is the deposited film surface. The resistivity, ρ can be calculate
using the Eq. (2.8), where Rf is the resistance of the prepared functional film.
(2.7)
(2.8)
25
(a)
(b)
Figure 2.6. Four terminals probe method to measure deposited film resistivity (a) schematic
diagram and (b) electrical circuit diagram.
26
2.5. Magnetic Field Analysis Simulation
The two-dimensional magnetic flux lines and their density profiles (Bx and Bz) are
analyzed by conventional Electromagnetic field analysis software poisson superfish – 7.18 [5]
developed at Los Alamos National Laboratory. It was analyzed by selecting the dimensions
and the magnetic properties of the permanent magnet. It was created an input coding that runs
on Poisson Superfish with TeraPad. It was analyzed by creating a mesh area by command
Automesh, and WSFplot command for display of mesh creation and confirmation. Magnetic
field analysis was performed by commanding Pandira. Output electronic text format data of
the magnetic field was made by the commanding SF 7. Using MS Excel text format data was
analyzed, and horizontal, Bx, and vertical, Bz magnetic field density profiles were evaluated.
2.6. Measurement of the Film Property
Atomic force microscopy (AFM) scanning was performed to observe the surface
morphology of the prepared films. In order to investigate the structure of the deposited film,
X-ray diffraction (XRD) standards (card no. ICDDPDF#04-0836) were used. Four-terminal
method was used the measure the film resistance of the prepared film.
27
References
[1] F. F. Chen, J. P. Chang, “Lecture Notes on Principles of Plasma Processing”, Plenum, New
York, 2002.
[2] M. A. Lieberman, and A. J. Lichtenberg, “Principles of Plasma Discharges and Material
Processing”, 2nd ed. John Wiley & Sons, Inc., New York, 2005.
[3] P. Chabert and N. Braithwaite, “Physics of Radio- Frequency Plasmas”, Cambridge
University Press, Cambridge, 2011.
[4] Y. Ohtsu, S. Tsuruta, T. Tabaru, and M. Akiyama, Surface Coatings Technology, 307,
1134–1138, 2016.
[5] http://laacg.lan.gov/laacg/services/serv_codes.phtml.
28
Chapter 3. Rotational Square-Shaped Setups of Rod Magnets for Uniform
Target Utilization
3.1. Introduction
Plasmas and their particles play an important role in nanotechnology and
semiconductor manufacturing [1-2]. Various kinds of plasma sources have been developed for
plasma processing. In particular, CCP driven by RF power supplies at 13.56MHz are widely
used for plasma processing, because their setup and maintenance are simple [3-4]. However,
CCP sources suffer from the fact that (1) productivity and deposition rates are low due to the
low plasma density of less than 109 cm-3 (depends on discharge conditions) and (2) the plasma
density and the energy of ions at the electrodes cannot be controlled separately in single
frequency CCPs [5]. On the other hand, it is well known that high-density plasmas can be
produced by hollow cathode discharges [6-18] because electrons are effectively confined in the
hollow trench.
The RF magnetized sputtering plasma source is widely utilized in the microelectronic
industry for functional thin film deposition [19-25] and to produce magnetic films [26-27]
because the setup is simple and their thin films are not conductive. Its application extends to
surface treatment and cleaning [13-14], flat panel display fabrication; transparent conductive
oxides film preparation [19-20] for solar cells and mobile phones and many other rapidly
growing areas [2]. In particular, the RF magnetron sputtering process has become popular for
the deposition of a variety of industrial surface coatings. This is because, at a RF input power
of more than 1000 W, the magnetron discharge plasma attains a high-density of charged
particles of 1010 - 1011 cm-3 at low gas pressures around 1 Pa due to plasma confinement by the
E×B drift motion [2], where E and B are the electric field perpendicular to the target and the
magnetic flux density parallel to the target, respectively.
The target surface is continuously eroded during the magnetron operation, but the
target material is not used effectively due to the non-uniformity of the plasma density. Thus,
the sputtering occurs highly localized in the region of the electron confinement. This leads to
the formation of a narrow and deep erosion groove and a low target material utilization [2]. In
conventional magnetron sputtering sources, the target utilization is approximately 20 to 30%
due to the narrow and deep groove erosion of the target [28]. It was reported that the target
erosion influences the ion distribution function in the near-cathode region as well as the
deposition rate to the substrate [29].
29
The estimated utilization of the target is increased to approximately 60-70% for planar
magnetron sputtering using a rotating tilted unbalanced and asymmetrical yoke magnet [30],
whereas RF magnetron plasmas based on stationary monopole arrangements attain a target
utilization of up to 59% for a magnet spacing of 10 mm [31]. Rotating the permanent magnet
in a circular cathode is a common technique for expanding the eroded area and increasing the
target utilization. However, at the center and outside the target, some non-eroded areas remain.
From the viewpoint of the practical operation of resources, improvement of the target material
utilization, uniform target erosion and high-density plasmas, as well as the prediction of
functional thin film deposition rates is required.
A high-density RF magnetized sputtering plasma source with a rotational square-
shaped magnet arrangement for uniform target utilization has been developed. This chapter
discusses a method to increase the target utilization and deposition rate for uniform erosion
over the entire area of the target. It is based on the rotation of a neodymium permanent rod
magnet of square-shaped arrangement with respect to the circular target. In section 3.2, the
experimental setup for the proposed RF magnetized plasma sputtering system is explained in
detail. In section 3.3.1, the magnetic field profiles obtained based on the squared-shaped
magnets are described. In section 3.3.2, the plasma emission and discharge characteristics are
discussed. In section 3.3.3, the copper target erosion profile and ion saturation current are
investigated explicitly. In section 3.3.4, the results obtained in this work are summarized.
3.2. Experimental Setup
The experiments are performed in a stainless-steel cylindrical vacuum chamber with
an outer diameter of 235 mm, an inner diameter of 160 mm and a height of 195 mm such as
shown in Fig. 3.1. In this experiment, a Cartesian coordinate system is used, because the
magnetic field is not axisymmetric. In this setup, before starting the experiment, an initial
vacuum pressure in the chamber of 2.8 10-4 Pa is realized based on a turbo molecular and an
oil rotary pumps. Argon (Ar) gas at 1.03 Pa is used by regulating a flow meter. An RF power
of 50 W at 13.56 MHz was applied to the target via an impedance matching network The
electrode (target) dimensions are 160×160×3 mm and the grounded wall. A copper plate with
3 mm thickness was used as the target to measure the sputtering characteristics based on
rotational square-shaped magnet schemes in the proposed high-density RF magnetized plasma
sputtering source. An aluminum disc was only used to measure plasma parameters in order to
avoid thin film deposition on the Langmuir probe. The discharge voltage between the RF
30
powered electrode and the grounded vacuum wall was measured by a high-voltage probe. The
typical discharge voltage changed from 800 to 600 Vpp (peak-to-peak value) with increasing
gas pressure from 1 Pa to 5 Pa at an RF input power of 50 W. The sputtering time period was
4 hours to measure the erosion depth of the target and to perform a calculation to obtain the
target utilization percentage. The origins of axial z and radial r positions are defined at the
surface and at the center of the RF powered target, respectively.
Eight Neodymium rectangular rod magnets of 30 mm length, 5 mm width, and 3 mm
height, where the connection between N-pole and S-pole magnets is one side of the squares,
are mounted on a circular iron yoke disk. The iron yoke is set up with a gap of 1 mm from the
copper target plate. An iron (Fe) cover of 5×3×1 mm is used for magnetic shielding of otiose
magnetic fields from the permanent magnets. The three magnet arrangements including
magnetic shielding material and the iron cover of 1 mm thickness are shown in Fig. 3.2 for the
proposed high-density RF plasma sputtering system. Details of the three magnet arrangements
will be explained later.
Figure 3.1. Experimental setup for the proposed high-density RF magnetized plasma
sputtering source.
31
The circular iron yoke disk with the magnet arrangement was rotated by an iron yoke
shaft at a speed of 40 rpm by an external motor. A precision surface tool was used to measure
the erosion depth of the copper target as a basis to calculate the target utilization percentage.
Experimental conditions are shown in Table – 3.1. The plasma parameters such as the time-
averaged ion saturation currents are measured to obtain the ion flux to the target by a tiny
tungsten wire probe of 1.0 mm in diameter and 10 mm in length. The time-averaged ion
saturation currents were measured by moving the L-shaped needle probe in axial z and radial r
directions as shown in Fig. 3.1. In order to avoid the influence of the RF potential oscillations
on the probe current-voltage characteristics, the probe wire was connected to an LC filter circuit
[1-5]. The ion flux to the target is calculated from the ion saturation current by negatively
biasing the probe including a resistance of 100 Ω positioned at z = 10 mm.
Table 3.1
Practical Experimental Conditions
Item Specification
Initial vacuum base pressure [Pa] 3.2 10-4
Argon gas introduce pressure [Pa) 1.03
Sputtering time [hours) 4
RF input power [W] 50
RF discharge [Vpp) 796
Self-biased DC voltage [V] -389
Rotational speed for iron yoke [rpm] 40
Neodymium magnet size [mm] 30 5 3
Copper target size [mm] 160 160 3
32
(a)
(b)
33
(c)
Figure 3.2 (a)-(c). Proposed square-shaped magnet arrangement (a) without iron (Fe) cover
and no air gap, (b) with iron (Fe) cover and no air gap and (c) with iron (Fe) cover and 5 mm
air gap. Here, the measured magnetic flux lines are also superimposed on these figures.
3.3. Results and Discussions
3.3.1. Magnetic Field Profile Analysis of Square-Shaped Schemes
In order to investigate the effect of the magnetic field pattern induced by the square-
shaped permanent rod magnet schemes on the plasma production, three arrangements of
magnets are used: (a) no iron (Fe) cover, no air gap between the N-pole and the S-pole magnets,
(b) with iron (Fe) cover, no air gap and (c) with iron (Fe) cover, 5 mm [see Fig. 3.2(a)-(c)].
Eight Neodymium rectangular rod magnets of 30 × 5 × 3 mm dimension are positioned in a
square-shape on a circular iron yoke disk of 140 mm in diameter. The iron yoke is set up with
a gap of 1 mm from the copper target plate. The surface magnetic flux density of the rectangular
rod magnets and the thickness of the iron yoke disc is 313 mT and 1 mm, respectively.
As shown in Fig. 3.2(a) - (c), the rectangular rod magnets are placed at a distance of
33 mm away from the central position. The measured magnetic flux lines parallel to the target
are superimposed on the magnet illustrations in Fig. 3.2. Figure 3.2 shows that the magnetic
flux lines are directed from the N-pole to the S-pole. As shown in Fig. 3.2 (a), the direction of
34
the electric field, E, and the behavior of the magnetic flux density, B, play an important role
for generating the magnetized discharge based on the E×B drift motion, where E and B are the
electric field perpendicular to the target and the magnetic flux density parallel to the target,
respectively. The direction of the E×B drift on the surface of the iron yoke disc is outward [see
Fig. 3.2(a)]. Four different E × B drift motions occur in the central region at each contact zone
between the N-pole and the S-pole magnets such as shown in Fig. 3.2 (a).
In this chapter, we propose to improve the uniformity of the target erosion by rotating
the magnets on the iron yoke. When the magnets are rotated at a fixed angular frequency, ,
the rotating speed is v = r, where r is the radial position. At the central region of the target,
the interactional time of ions to the target becomes longer, while for the edge region the
interactional time becomes shorter. This inward drift motion increases the plasma density in
the central region of the target so that the erosion depth is higher than that in the edge region
of the target. Thus, the suppression of the inward drift motion is important. In order to realize
a uniform target erosion, the inward drift motion should be suppressed. The E×B drift motion
can be prohibited by a magnetic shielding material, i.e. an iron cover of 1 mm thickness. The
iron cover of 5 mm in length, 3 mm in width and 1 mm in thickness was used for the magnetic
shielding such as shown in Figs. 3.2 (b) and (c). The influence of the iron cover mounted on
the contact zone between the N-pole and the S-pole magnets [see Fig. 3.2 (b)] and the air gap
between the N-pole and the S-pole magnets [see Fig. 3.2 (c)] on the magnetic field profile are
investigated to suppress the magnetic flux density which enhances the inward drift motion of
the plasma particles.
The two-dimensional magnetic flux lines and their profiles of one–pair magnet in the
square-shaped magnet scheme are analyzed by the conventional magnetic-field analysis
software Poisson Superfish developed in Los Alamos National Laboratory [32]. Figure 3.3
shows the two-dimensional distributions of the magnetic flux lines near the target surface for
cases (a), (b) and (c), respectively. For all square-shaped magnet arrangements and based on
the two-dimensional distributions of the magnetic flux lines near the target surface, the
magnetic field lines starting at the N-pole of the magnets pass through the target, airspace, iron
yoke and then return to the S-pole.
35
(a)
(b)
36
(c)
Figure 3.3. Two-dimensional distributions of magnetic flux lines near the target surface (a)
without iron (Fe) cover and no air gap, (b) with iron (Fe) cover and no air gap, and (c) with
iron (Fe) cover and 5 mm air gap.
Figure 3.4 shows radial distributions of the absolute value of the horizontal magnetic
flux density Bx for the magnet arrangements of cases (a), (b), and (c) at the axial position z
=10 mm. The shape of their peaks is almost the same for all square-shaped magnet
arrangements and the radial profile of the magnetic flux density has three peaks. The
amplitudes of two peaks positioned at x = 3.5 and 3.5 cm are similar for all magnet
arrangements because their position is away from the iron cover. The peak amplitude of |Bx|
for case (b) is the lowest among all magnet arrangements. The magnetic shielding material, i.e.
the iron cover, reduces the horizontal magnetic flux density at x = 0 from approximately 65 to
50 mT. On the other hand, the peak amplitude of |Bx| with the iron cover at r = 0 increases as
a function of the air gap between the magnets. Thus, the cover effect is useful to decrease the
E×B drift towards the center of the target. Thus, the uniformity of the copper erosion profile
will be improved by the presence of the iron cover. The copper target utilization for case (b) is
maximum. From the magnet axis, the direction of the magnetic flux density |Bx| is reversed on
the left and right side such as shown in Fig. 3.4. The ring-shaped plasma formed around the
magnet drifts in opposite directions. These two ring discharges interact with each other. This
37
interaction determines the formation and the shape of the final ring discharge as well as the
position of its center.
Figure 3.5 shows the radial distributions of the absolute value of the axial component
|Bz| the magnet arrangements of cases (a), (b) and (c) at the axial position z =10 mm. It is found
that |Bz| has two peaks and that, based on Fig. 3.5, case (b) corresponds to the lowest amplitude
whereas case (a) corresponds to the highest magnitude among all square-shape magnet schemes.
Thus, case (b) is the best arrangement, because the axial component does not play a role in the
E B drift motion.
Fig. 3.4. Horizontal magnetic flux density as a function of x position at z = 10 mm.
38
Figure 3.5. Vertical magnetic flux density as a function of x position at z = 10 mm.
3.3.2. Plasma Emissions and Discharge Characteristics
Figure 3.6 shows the plasma emission for case (b) at an RF power of 50 W. The plasma
emission was measured by a low-resolution digital camera. The yellow and white colored
rectangles mark the positions of the permanent rod magnets. Strong emission is observed in an
outward direction at the four intersections between the four vertically placed magnets. As
mentioned in the magnetic field section, the plasma emission in outward direction is caused by
the E B drift motion effect. It is also found that a square-shaped plasma is observed in the
center region. At the top of the square-shaped plasma, plasma emission towards the center is
observed. The square-shaped plasma is formed by the effect of the magnetic field resulting
from each magnet. It seems that the plasma density at the edge of the target is higher than that
at the center. For the other magnet schemes such as cases (a) and (c), the plasma emission
images are similar at an argon gas pressure of 1 Pa and a RF power of 50 W. As described in
the magnetic field analysis section, for case (b) we also observe that four E × B drift motions
in the central direction are suppressed compared with the other cases. These inward drift
motions as well as the plasma density at the central direction of the target are prohibited by the
magnetic shielding material, i.e. the iron cover of 5×3×1 mm dimension, in order to achieve a
uniform target erosion.
39
Figure 3.6. Plasma emissions for (b) iron (Fe) cover, no air gap setup at RF power of 50 W.
The RF discharge voltage between the RF powered electrode and the grounded
chamber wall is measured by a digital oscilloscope and a high-voltage probe. The RF power is
fixed at 50 W. Argon gas is introduced into the chamber at 1 to 5 Pa. Fig. 3.7 shows the RF
discharge voltage VRF (peak-to-peak value) as a function of the Ar gas pressure for case (a),
(b), and (c). It is seen that the RF discharge voltage decreases gradually from approximately
760 to 600 Vp-p with increasing the argon gas pressure for all cases. This is because the electron
collision frequency increases with increasing Ar gas pressure so that the ionization rate
increases. According to the absorbed power relation, where ohmic heating is assumed in the
sheath [2], at a fixed absorbed power, the RF discharge voltage is roughly inversely
proportional to the electron-collision frequency.
The absolute value of the dc self-bias voltage, |Vsb|, of the cathode is shown in Fig. 3.8
for all square-shaped magnet schemes. The self-bias voltage of the cathode is negative. Its
absolute value decreases from approximately 400 to 270 V with increasing the gas pressure for
all cases. This tendency is attributed to the decrease in the RF discharge voltage between the
cathode and the grounded chamber wall. It is found from Fig. 3.7 and Fig. 3.8 that the case (a)
40
has the lowest amplitude and the case (c) has the highest magnitude among all square-shape
magnet schemes. These results are similar for the copper target erosion depth profile, which
will be discussed later. For the sputtering of a copper target, the Ar gas pressure of 1 Pa is
selected as a conventional sputtering gas pressure.
Figure 3.7. RF Discharge voltage between the cathode and grounded chamber wall as a
function of argon gas pressure.
41
Figure 3.8. DC self-bias voltage on the cathode as a function of argon gas pressure.
3.3.3. Profiles of the Target Erosion Depth and Ion Flux to the Cu target
The copper sputtering experiments were done by rotating the iron yoke disk at a speed
of 40 rpm for the square-shaped magnet arrangements of cases (a) - (c). Figure 3.9 shows the
radial profiles of the copper target erosion depth for the different magnetic arrangements at an
RF power of 50W and a processing time of 4 hours. As shown in Fig. 3.9, the highest target
erosion depth is observed at r = 30 mm for cases (a) and (b), whereas for case (c), the highest
target erosion depth is found at r = 25 mm. The utilization percentage of the copper target was
estimated based on the ratio of the copper target erosion volume to the volume calculated under
the assumption that the target is completely eroded at a maximum erosion depth. The target
utilization percentages are 74.15, 87.49 and 83.85 % for case (a), (b) and (c), respectively. For
the case (b), the copper target utilization percentage reaches the highest value of 87.49% among
all square-shaped magnet arrangement schemes.
It is seen that the copper target erosion homogeneity is increased by decreasing the air
gap compared with cases (b) and (c) in Fig. 3.9. The magnetic shielding characteristic has been
investigated in order to realize a uniform target erosion depth. The copper target utilization
percentage is increased from 74.15% to 87.49% by using the magnetic shielding iron (Fe) cover
of 1 mm thickness. Moreover, by decreasing the air gap between the shielded magnets, the
copper target utilization percentage is increased from 83.85% to 87.49%. However, a high
42
target erosion depth with some fluctuation and non-uniformity is observed for the case (c) in
Fig. 3.9. The uniformity of the target erosion depth has been improved by using the iron (Fe)
cover and decreasing the air gap between the shielded magnets. Thus, the iron (Fe) cover is
effective to prohibit an inward E×B drift motion to the center of the target so that the uniformity
of the copper erosion profile as well as the target utilization is improved by the addition of the
iron cover and reducing the air gap between one–pair shielded magnet.
In order to study the ion flux to the target, the ion saturation current is measured as a
function of the radial position by a Langmuir probe. Aluminum disks were used to avoid the
deposition of thin films to the probe for this measurement. As shown in Fig. 3.6, since the
plasma distribution is not axially symmetry, the ion current detected by the probe changes
temporally when the magnets are rotating. Temporal variations of the ion saturation current
were measured at each radial position from r = 0 to 45 mm. The time-averaged ion saturation
current is estimated from these temporal variations.
Figure 3.9. Radial profiles of the copper target erosion depth at RF power of 50 W.
43
Figure 3.10 shows the radial profiles of the time-averaged ion saturation current, Iisat
for cases (a) - (c). The ion saturation current is measured by a tiny probe biased at Vp = - 70 V
at z = 10 mm. The RF power was fixed at 50 W. According to the corresponding probe theory
[1], the ion saturation current is proportional to the ion flux to the target. It is shown that the
uniformity of the ion saturation current for case (b) is better than for the other square-shaped
magnet arrangement. For case (a), the time-averaged ion saturation current is lower than for
the other proposed scheme as shown in Fig. 3.10. For case (b), the radial profile of the ion flux
to the target is approximately similar to the radial profile of the copper target erosion depth as
shown in Fig. 3.9. The sputtering rate Rsp of the copper target is expressed using the following
Eq. (3.1), where is the yield of sputtered atoms per incident ion. , and e are the area
of the probe used for detecting the time-averaged ion saturation current, the atomic density of
the copper target and the electronic charge, respectively.
(3.1),
Therefore, the erosion depth profile is consistent with the ion flux profile. The plasma
density is calculated from the time-averaged saturation current for cases (a) - (c) using the
Bohm sheath criteria [1]. Measurements were performed at a position of x = 0 mm and z =10
mm. The calculated plasma densities are 8.8×109 cm-3, 1.1×1010 cm-3, 1.0×109 cm-3 for cases
(a) - (c), respectively. The plasma density was the highest for case (b) among all square shaped
arrangements.
44
Figure 3.10. Time-averaged ion saturation current as a function of the radial position detected
by Langmuir probe at various square-shaped magnet arrangements.
45
3.4. Conclusion
A square-shaped magnet arrangement consisting of eight neodymium rod magnets of
30 × 5 × 3 mm has been investigated in order to realize a uniform utilization of a copper target
in an RF magnetized sputtering plasma. A copper plate of 160 × 160 × 3 mm and an aluminum
disc of 160 mm in diameter were used as the target to measure the sputtering characteristics
based on rotational square-shaped magnet schemes in the high-density RF magnetized plasma
sputtering source. The magnetic shielding material, iron (Fe) cover, suppresses the horizontal
magnetic flux density |Bx| at x = 0 from approximately 65 to 50 mT. It is also found that the
case (b) with iron (Fe) cover and no air gap has the lowest amplitude of the horizontal magnetic
flux density among all square-shape magnet schemes. Strong plasma emission in an outward
direction is observed at four intersections between the four vertically placed magnets. It is seen
that the RF discharge voltage and the absolute value of the dc self-bias voltage decrease
gradually from approximately 760 to 600 Vp-p and approximately 400 to 270 V, respectively,
with increasing argon gas pressure for all cases. As the sputtering rate is determined by the ion
flux to the target, it is found that the target erosion profile approximately agrees with the ion
flux profile for case (b). The uniformity of the target erosion depth and ion flux profile has
been improved by using the iron (Fe) cover and decreasing the air gap between the shielded
magnets. The iron (Fe) cover is effective to prohibit the inward E×B drift motion to the center
of the target so that the uniformity of the copper erosion profile is improved by the addition of
the iron cover and reducing the air gap between the one–pair shielded magnet and the target
utilization reaches the highest value of 87.49%.
46
References
[1] F. F. Chen, J. P. Chang, Lecture Notes on Principles of Plasma Processing, Plenum, New
York, 2002, [part A1, B1, A7].
[2] M. A. Lieberman, A. J. Lichtenberg, Principles of Plasma Discharges and Material
Processing, 2nd ed. John Wiley & Sons, Inc., New York, 2005, [chapter 1, chapter 4].
[3] Y. Ohtsu, Y. Yahata, J. Kagami, Y. Kawashimo, T. Takeuchi, IEEE Trans. Plasma Sci., 41,
1856 - 1862., 2013.
[4] Y. Ohtsu, N. Matsumoto, J. Vac. Sci. Technol. A, 32, 031304, 2014.
[5] Y. Ohtsu, H. Urasaki, Plasma Sources Sci. Technol., 19, 045012, 2010.
[6] S. Hashiguchi, M. Hasikuni, Jpn. J. Appl. Phys., 26, 271-280, 1987.
[7] T. Lafleur, R. W. Boswell, Phys. Plasma, 19, 023508, 2012.
[8] H. Pak, M. J. Kushner, J. Appl. Phys. 71, 94 – 100, 1992.
[9] H. Eichhorn, K. H. Schoenbach T. Tessnow, Appl. Phys. Lett., 63, 2481 – 2483, 1993.
[10] A. R. Petre, M. Bazavan, V. Covlea, V. V. Covlea, A. H. Oprea, Rom. Rep. Phys., 56, 271
– 276, 2004.
[11] M. Sugawara, T. Asami, Surf. Coating Technol., 74-75, 355 -357, 1995.
[12] T. Fukuda, A. Matsuoka, Y. Kondoh, M. Sugawara, Jpn. J. Appl. Phys., 37, L81-L64,
1998.
[13] H. Barankova, L. Bardos, Surface Coating Technol., 146-147, 486 – 490, 2001.
[14] L. Bardos, H. Barankova, Y. A. Lebedev, Surface Coating Technol., 163-164, 654 -658,
2003.
[15] T. Tabuchi, H. Mizukami, M. Takashiri, J. Vac. Sci. Technol. A, 22, 2139-2144, 2004.
[16] K. Yambe, A. Matsuoka, Y. Kondoh, Jpn. J. Appl. Phys., 45, 8883-8889, 2006.
[17] Y. Ohtsu, H. Fujita, Appl. Phys. Lett., 92, 171501-171503, 2008.
[18] Y. Ohtsu, Y. Kawasaki, J. Appl. Phys., 113, 033302, 2013.
[19] S. Rahmane, M. S. Aida M. A. Djouadi, N. Barreau, Superlattices and Microstructures,
79, 148-155, 2015.
[20] A. Spadoni, M. L. Addonizio, Thin Solid Films, 589, 514-520, 2015.
[21] J. Kohout, E. Bousser, T. Schmitt, R. Vernhes, O. Zabeida, J. K. Sapieha, L. Martinu,
Vacuum, 124, 96-100, 2016.
[22] H. M. R. Giannetta, C. Calaza, D. G. Lamas, L. Fonseca, L. Fraigi, Thin Solid Films, 589,
730-734, 2015.
[23] C. Anil Kumar, D. Pamu, Applied Surface Science, 340, 56-63, 2015.
47
[24] A. I. Savchuk, V. V. Strebezhev, G. I. Kleto, Y. B. Khalavka, I. M. Yuriychuk, P. M.
Fochuk, Surface Coating Technol., 295, 8-12, 2015.
[25] B. Xu, X. G. Ren, G. R. Gu, L. L. Lan, B. J. Wu, Superlattices and Microstructure, 89, 34-
42, 2016.
[26] S. Gider, B. U. Runge, A. C. Marley, S. S. P. Parkin, Science, 281, 797-799, 1998.
[27] W-G. Wang, M. Li, S. Hageman, C. L. Chien, Nature Materials, 11, 64-68, 2012.
[28] T. Van Vorous, Solid State Technol., 19, 62, 1976.
[29] M. Ganeva, A. V. Pipa, R. Hippler, Surface Coating Technol., 213, 41-47, 2012.
[30] T. Iseki, Vacuum, 84, 339 -347, 2010.
[31] Y. Ohtsu, M. Shigyo, M. Akiyama, T. Tabaru, Vacuum, 101, 403 – 407, 2014.
[32] http://laacg.lan.gov/laacg/services/serv_codes.phtml.
48
Chapter 4. Performance of a Gyratory Square-Shaped Capacitive Radio
Frequency Discharge Plasma Sputtering Source for Material Processing
4.1. Introduction
Plasma processing has recently drawn significant attention owing to its promising
applications, including in semiconductor and manufacturing industries around the world [1-2].
Many plasma sources and experiments developed in the past four decades have been discussed
for the purposes material processing and functional film preparation [2-3]. CCPs functioned by
a RF power supply at 13.56 MHz operating in various gases are, in principle, some of the
simplest, and easily maintained methods of large-diameter substrate processing, and are
important in many high-tech applications, including etching, deposition of thin films,
modification of surface properties and microelectronic device fabrication [4–8]. In particular,
CCPs have some limitations, such as a low plasma density of less than 109–1010 cm-3, a low
deposition rate, the standing wave effect [8] on the powered electrode (target), as well as the
difficulty of controlling external parameters and energized ions at the target independently [4–
9]. Therefore, the conventional CCPs are not a tool suitable for plasma processing. The physics
and chemistry of capacitively coupled RF plasmas have been the subject of many theoretical
and experimental studies via modeling; in order to overcome the limitations, a number of
solutions have been employed [1, 10–12]. Some solutions include very high-frequency (VHF)
operation to increase the plasma density at a given input power [1, 8, 12], optimization of the
gas mixing to maximize the ionization probability [14], and the use of the high-secondary
electron emission electrodes [6, 13]. Two mechanisms [15-16] of electron heating, such as (1)
ohmic heating due to collisions of electrons with neutral gases and (2) stochastic heating—
often referred to as Fermi heating—due to momentum transfer from the oscillating RF sheath,
pressure and ambipolar electron heating [17], as well as voltage waveform tailoring [18], play
a major role in capacitively coupled RF discharges. Different methods [19–20] have been
proposed in order to study electron heating phenomena in capacitive discharges. Experiments
have also been performed to investigate stochastic heating [21–23]. Many authors have
investigated the characteristics of dual-frequency (DF) inductively coupled plasma (ICP) [24],
CCP discharges, experimentally [25–26]. Moreover, a novel type of DF-CCP operating at
consecutive harmonics instead of at substantially different driving frequencies was proposed
based on the electrical asymmetry effect (EAE), in order to avoid the limitations of the
independent control of sputtering parameters, such as the density and energy of ions incident
on the target [27]. In general, it is well known that hollow-cathode discharges [9, 28–32] can
49
produce high-density plasma using only a simple structure for the hollow electrode because a
hollow trench can confine electrons effectively. Recently, RF magnetized plasma sources have
been widely used in microelectronics [33–35], such as magnetic films [36], surface treatment
and cleaning [30], diamond-like carbon [37], biomaterial thin films, flat panel display
fabrication, transparent conductive oxide film preparation [33, 35] for solar cells and mobile
phones, and many other rapidly growing areas [1]. In particular, RF magnetron plasma sources
have become an attractive tool for functional film preparation. However, the conventional
magnetron plasma has an issue that the target erosion is not uniform owing to high-density
localized plasmas.
In our previous work, a high-density RF magnetized sputtering plasma source with a
rotational square-shaped magnet arrangement was developed for uniform target utilization
about 88 % uniformity [38]. The results revealed that the uniformity of the target erosion depth
and ion flux profile is improved by using magnetic shielding cover made from iron (Fe). It was
also found that the iron cover suppresses the horizontal magnetic flux density and effectively
prohibits the inward E×B drift motion towards the center of the target [38]. Copper films are
fabricated using different techniques, such as DC magnetron sputtering, pulsed laser deposition
(PLD), RF magnetron sputtering, ion beam assisted DC magnetron reactive sputtering, and
molecular beam epitaxy (MBE) [39-41]. The RF magnetized plasma sputtering method is
selected because it has many advantages, such as the possibility of large area deposition, a
simple setup arrangement, and uniform deposition capabilities.
The objectives of this paper are to investigate the performance of a gyratory square-
shaped RF magnetized plasma sputtering source for materials processing and different
functional film preparations. For example, Cu films have also been deposited on Si wafer
substrates by our developed RF magnetized sputtering plasma source at an RF power of 100
W with a processing time of 1.5 h. A radial profile and microstructure of the deposited films
are investigated under two square-shaped magnet schemes. In Section 4.2, the experimental
setups are explained in detail. In Section 4.3, the results and discussions on the effects of the
magnetic shielding material in a square-shaped magnet arrangement on the discharge
characteristics, plasma emission luminescence, thickness, and resistivity profile, as well as the
surface morphology of the film deposited by the gyratory square-shaped RF magnetized plasma
sputtering source are presented. In Section 4.4, the results obtained in this chapter are
summarized.
50
4.2. Experimental Setup Details
The schematic of the gyratory square-shaped capacitive RF discharge plasma
sputtering source for plasma processing is shown in Fig. 4.1. The experiments were performed
in a stainless-steel cylindrical vacuum chamber with an outer diameter of 235 mm, an inner
diameter of 160 mm and a height of 195 mm. Eight Neodymium rectangular permanent bar
magnets with a volume of 30 mm × 5 mm × 3 mm are mounted on a circular iron yoke disk so
that the line joining the N-pole and S-pole magnets comprises one axis of the square. Figure
4.2 shows this one axis of the proposed square-shaped magnet scheme for case (a), without
iron shielding, and case (b), with iron shielding of dimensions 5 mm × 3 mm × 1 mm. The
chamber was evacuated by a combination of a turbo molecular and an oil rotary pump, and the
base pressure in the chamber of 1.6 10-5 Pa is obtained before starting the experiment. Argon
gas was introduced into the vacuum chamber as the sputtering at a working gas pressure of
1.02 Pa through a regulating flow meter. The iron yoke was set up at a distance of 1 mm from
the target surface. An iron cover of 5 mm × 3 mm× 1 mm is used for magnetic shielding in
order to suppress the magnetic flux lines between the contact zone of the N-pole and S-pole
magnets, as shown in Fig. 4.2. The circular iron yoke disk mounted by the magnets was rotated
at a speed of 40 rpm by an external motor drive. A cooling fan was used to cool the magnets.
An RF power of 100 W at 13.56 MHz was applied to the electrode (target) via an impedance-
matching network and a blocking capacitor, while the vacuum chamber wall acted as a
grounded electrode. An aluminum disc target was used, and the substrate holder and Si wafer
substrate shown in Fig. 4.1 were also removed to measure the plasma emission image and
discharge characteristics in order to avoid film deposition on the glass window at the bottom.
A copper plate of 160 mm × 160 mm × 3 mm was used as the target to deposit a copper film
on the Si wafer substrate in the proposed gyratory square-shaped magnet arrangements in the
RF magnetized plasma sputtering source. The Si substrate was ultrasonically cleaned in
distilled water for 7 min and water drops were removed before inserting it into the deposition
chamber. The Si substrate was fixed onto the substrate holder and the distance between the
target and the substrate was 45 mm. A precision surface profile tool was used to measure the
radial profile of the thickness of the deposited Cu film.
51
Figure 4.1. Schematic of the gyratory square-shaped RF magnetized plasma sputtering
source set up for materials processing.
Figure 4.2. One axis of the proposed square-shaped magnet scheme for case (a), without iron
shielding, and case (b), with iron shielding of dimensions 5 mm × 3 mm × 1 mm.
The resistivity of the deposited Cu films was measured by a standard four-point probes
technique at room temperature. The surface morphology of the Cu films was observed using
AFM. The thin film deposition processing time period was 1.5 h. The RF discharge voltage
between the RF powered electrode and the grounded vacuum wall, as well as a DC self-biased
voltage, was measured by a high-voltage probe with an attenuator of high input impedance and
a digital oscilloscope with a sampling frequency of 2 Giga sampling/sec. As shown in Fig. 4.1,
the origins of the axial z and radial r positions are defined at the surface and at the center of the
RF powered target, respectively. The plasma density is estimated from the ion saturation
current detected by a negatively biased [4] cylindrical tungsten probe of 1.0 mm in diameter
and 10 mm in length. The estimation of plasma density from the ion saturation current is done
52
by moving the L-shaped probe to the radial direction, as shown in Fig. 4.1, and measurements
were taken at r = 5 mm and z = 10 mm. The probe wire was compensated with an LC filter
circuit in order to minimize the influence of the RF plasma potential oscillations on the probe
current-voltage characteristics [7]. A Cartesian coordinate system is used because the magnetic
field is not axisymmetric. The typical experimental conditions used to investigate the
performance of a gyratory square-shaped capacitive RF discharge plasma sputtering source for
materials processing are shown in Table 4.1.
Table 4.1
Typical experimental conditions of a proposed plasma
sputtering source for materials processing.
Parameters Specification
Initial base pressure [Pa] 1.6 10-5
Target electrode Al/Cu
Substrate Si wafer
Sputtering gas Ar
Argon gas pressure [Pa] 1.02
RF input power [W] 100
Rotational speed of yoke [rpm] 40
Sputtering time [h] 1.5
Distance between the target and rotatory
yoke with magnets [mm] 1
Target - Substrate distance [mm] 45
RF discharge voltage [Vpp] 1220
DC self-biased voltage [V] - 608
The plasma discharge image is taken by a low-resolution digital camera. A typical
plasma discharge image with for case (b) at an RF power of 100 W is shown in Fig. 4.3, and
also clarifies the magnet arrangement. The positions of the permanent rod magnets are
indicated by the red and blue colored dashed rectangles. The gray colored squares represent the
iron cover. Eight Neodymium rectangular permanent bar magnets of 30 mm × 5 mm × 3 mm
are mounted on a circular iron yoke disk so that the line joining the N-pole and S-pole magnets
53
comprised one axis of the square. Strong emissions of the plasma discharge are observed in the
outward direction at the four intersections of the four vertically placed magnets and in the
inward direction at the four connections between N-pole and S-pole magnets, which are marked
by eight arrows. According to right-hand rule, the plasma discharge in the inward and outward
directions is caused by the E B drift motion effect.
Figure 4.3. A typical plasma discharge image for case (b) at an RF power of 100 W, a
gyratory speed of 40 rpm, and Ar gas pressure of 1 Pa. Eight arrows denote the directions of
E×B drift motion.
54
4.3. Results and Discussions
4.3.1. Magnetic Shielding Effect Square-Shaped Magnet Arrangement
The horizontal and the vertical magnetic flux densities, and their radial profiles with
one–pair magnets as one axis of the square-shaped permanent rod magnet scheme, are analyzed
by the conventional magnetic-field analysis software, Poisson Superfish [42], in order to
investigate the effect of the magnetic shielding on the plasma discharge, thickness and
resistivity profiles of deposited copper thin film. Two magnetic arrangements were used: case
(a) without iron shielding and case (b) with iron shielding of 5 mm × 3 mm × 1 mm iron
material. As shown in Fig. 4.3, eight Neodymium rectangular rod magnets of 30 mm × 5 mm
× 3 mm in dimensions are placed at a distance of 33 mm away from the central position in a
square-shape on a circular iron yoke disk of 140 mm in diameter and 1 mm in thickness.
However, the Cu target covers an effective radial area from r = −50 mm to r = +50 mm inside
the vacuum chamber because the inner diameter of the hole of a circular insulator is 100 mm.
The surface magnetic flux density of the rectangular rod magnets of 3130 G was used. The
direction of the electric field, E, and the parallel component of the magnetic flux density, B,
are considered to significant for the magnetized plasma discharge based on the E×B drift
motion, where the electric field, E, and the magnetic flux density, B, are perpendicular to the
target and parallel to the target, respectively.
It was shown in our previous work [38] that four inward and outward E × B drift
motions occur in the central region at each contact zone between the N-pole and the S-pole
magnets and on the surface of the iron disc, respectively. The suppression of these inward drift
motions was necessary, because inward drift motion increases the plasma density in the central
region of the target so that the erosion depth was higher than that in the edge region of the
target. In order to deposit uniform film, these inward drift motion should be suppressed. These
four inward E×B drift motions can be prevented by a magnetic shielding material, the iron
cover. The iron covers positioned on the contact zone between the N-pole and the S-pole
magnets, such as the magnetic field profile, are investigated to suppress the magnetic flux
density that enhances the inward drift motion of the plasma particles. In our previous work [38],
it was shown that, near the target surface, in the distributions of the two-dimensional magnetic
flux lines, the magnetic field lines starting at the N-pole of the magnets pass through the target,
airspace, iron yoke and then return to the S-pole for both cases. Figures 4.4 and 4.5, respectively,
show the distributions of the absolute value of the horizontal, Bx, and the vertical, |Bz|,
magnetic flux densities for the magnet arrangements of cases (a) and (b), at the axial position
55
z = 5 mm from the copper target. It is observed that |Bx| and |Bz| have three and two peaks,
respectively for both cases. The direction of the magnetic flux density, |Bx,| is reversed on the
left and right sides of the magnet axis, as shown in Fig. 4.4. The peak amplitude of |Bx| for
case (b) is lower than that for case (a) and the amplitudes of the three peaks are located at x =
35, 0 and 35 mm, where x is the center of the target. The peaks of |Bz| are positioned at x =
45, 15, 45 and 15 mm. The magnetic shielding material reduces the horizontal magnetic flux
density at x = 0 from approximately 1120 to 824 Gauss and the axial component from
approximately 824 to 706 Gauss at x = 15 and 15 mm.
Figure 4.4. Distributions of the absolute value of the horizontal, Bx, magnetic flux densities
for the magnet arrangements of cases (a) and (b), at the axial distance z = 5 mm from the
target.
56
Figure 4.5. Distributions of the absolute value of the vertical, |Bz|, magnetic flux densities
for the magnet arrangements of cases (a) and (b), at the axial distance z = 5 mm from the
target.
The cyclotron angular frequency and larmor radii of charged particles such as
electrons and ions are important parameters for magnetized plasma. These cyclotron
frequencies decrease in a magnetic field reduced by the shielding material of the iron cover
according to the equation: , where q and m are the electronic charges and masses of
charged particles, respectively. Moreover, the larmor radius, rL, with the shielding material,
will be larger than those without the shielding material in accordance with , where v is
the particle movement speed perpendicular to magnetic field, B. Therefore, the magnetic
shielding effect caused by the iron cover is suitable for decreasing the E×B drift of ions toward
the center of the target so that the profile of the sputtered copper atoms will be uniform. The
uniformity of the radial profile of the thickness and the resistivity of the deposited Cu film will
be improved by the use of the shielding material in a gyratory square-shaped RF magnetized
plasma sputtering source. Consequently, case (b) will be the effective arrangement for
functional film preparation and material processing.
57
4.3.2. Plasma Discharge Voltage and Emission for Square-Shaped Magnetized Plasma
The RF discharge voltage, VRF, between the RF powered electrode and the grounded
chamber wall, and the self-bias dc voltage, Vsb, of the electrode are measured by a digital
oscilloscope and a high-voltage probe at various Ar gas pressures. Figures 4.6 and 4.7 show
the discharge voltage, VRF, and absolute value of the DC self-biased voltage, |Vsb|, respectively,
of the gyratory square-shaped RF magnetized plasma sputtering source as a function of the
sputtering gas pressure for cases (a) and (b) at a gyratory speed of 40 rpm and an RF power of
100 W. The RF discharge voltages, VRF, were 1160, 1125, and 1100 Vp-p and the absolute value
self-bias dc voltages, |Vsb|, were 573, 548, and 532 V with the argon gas pressures of 1, 2 and
3 Pa, respectively, for case (b) at an RF power of 100 W. These decreasing tendency results
are compatible with our previous work [38] at an RF power of 50 W. This is because the
electron collision frequency, as well as the ionization rate, increases with an increasing Ar gas
pressure, and the discharge voltage is approximately inversely proportional to the electron
collision frequency at a fixed absorbed power [1–2]. From the time-averaged ion saturation
current, the plasma density was also estimated using the Bohm sheath criteria [1–2] for both
cases. The measurements were positioned at r = 5 mm and z = 10 mm. The estimated plasma
densities were 8.4 × 1015 m-3 and 1.1 × 1016 m-3 for cases (a) and (b), respectively.
In the sputtering deposition, the RF power and sputtering Ar pressure were fixed at 100
W and 1 Pa, respectively. The RF discharge voltages, VRF, (peak-to-peak value) were 1080 and
1160 Vp-p for cases (a) and (b), respectively. The absolute values of the self-bias dc voltages,
Vsb, of the electrode were 528 and 573 V for cases (a) and (b), respectively. A typical star-
shaped plasma is also observed for both square-shaped cases. The star-shaped plasma is formed
by the effect of the magnetic field produced by each permanent rod magnet. As described in
the magnetic shielding effect section, four E × B drift motions in the central direction are
suppressed by the iron cover. In order to deposit a uniform Cu film, these inward drift motions,
as well as the plasma density, toward the central point of the target are prohibited by the
magnetic shielding material of the iron cover of 5 mm × 3 mm × 1 mm in dimension.
58
Figure 4.6. Discharge voltage, VRF, of the gyratory square-shaped RF magnetized plasma
sputtering source as a function of the sputtering gas pressure for cases (a) and (b), at a
gyratory speed of 40 rpm and an RF power of 100 W.
Figure 4.7. The absolute value of the DC self-biased voltage, |Vsb|, of the gyratory square-
shaped RF magnetized plasma sputtering source as a function of the sputtering gas pressure
for cases (a) and (b), at a gyratory speed of 40 rpm and an RF power of 100 W.
59
4.3.3. Thickness and Resistivity Profiles of the Deposited Film of the Gyratory Square-
Shaped RF Magnetized Plasma Sputtering Source
The performance of the gyratory square-shaped RF magnetized plasma sputtering
source was analyzed for materials processing on the microstructure of deposited films. Figure
4.8 shows the film thickness as a function of the radial position for cases (a) and (b) at an RF
power of 100 W, a sputtering gas pressure of 1.02 Pa, a gyratory speed of 40 rpm, and a
processing time of 1.5 h. We have shown in our previous work [38] that the copper target
utilization percentage is increased from 74.15% to 87.49% by using a magnetic shielding iron
cover of 1 mm thickness. As shown in Fig. 4.8, the highest and lowest thicknesses of the copper
thin film are observed at r = 30 and r = 40 mm for both cases, respectively. In Fig. 4.8, it can
be observed that the uniformity of the deposited film thickness is obviously improved by using
the iron shielding material on four paired magnets. The thickness of the deposited film for case
(b) is around 0.6 m for 0 < r < 30 mm and then decreases at the edge of the target, whereas
the thickness profile is fluctuated significantly for case (a). The roughness is estimated based
on the ratio of the highest thickness minus the lowest thickness and the highest thickness plus
the lowest thickness of the film. The roughnesses of the thickness profiles of the films are
approximately 24.4 % and 7.2 % for case (a) and (b), respectively. It is considered that the
thickness profile of the thin film is affected by the copper target erosion depth profile. These
findings are consistent with those reported in our previous work [38], in which the uniformity
of the copper target erosion depth profile was improved by the use of the shielding material,
and which predicts the uniformity of the copper thin film thickness profile.
60
Figure 4.8. Film thickness as a function of the radial position for cases (a) and (b), at an RF
power of 100 W, a sputtering gas pressure of 1.02 Pa, a gyratory speed of 40 rpm, and a
processing time of 1.5 h.
In order to evaluate the electrical properties of the deposited film at an RF power of
100 W with a sputtering duration of 1.5 h, the resistance and the resistivity measurements are
carried out at room temperature by the four-point probe method. The resistivity of the film
under room temperature as a function of the radial position for cases (a) and (b), where films
were deposited at an RF power of 100 W, a gyratory speed of 40 rpm, a sputtering gas pressure
1.02 Pa, and a processing time of 1.5 h, is shown in Fig. 4.9. It can be seen that the resistivity
profile of case (b) is more uniform than that of case (a). The lowest electrical resistivity of the
deposited film is 4.33 × 10-8 m at r = 30 mm for case (b), which is virtually the same as the
bulk resistivity of the copper. The resistivity decreases from 9.46 × 10-8 to 4.69 × 10-8 m at
r = 20 mm when using the shielding material. The resistivities are 8.16 × 10-8 and 4.64 × 10-8
m at r = 0 mm for cases (a) and (b), respectively.
The film property can be affected by the energy of the sputtered copper atoms incident
on the film. The energy of Cu atoms depends on the energy of Ar ions incident on the Cu target,
which is determined by the discharge voltage, that is, the self-biased voltage on the target. As
mentioned in above, the discharge voltage and the self-biased voltage for case (a) are almost
the same as those for case (b). In this deposition, the difference in the resistivity between cases
(a) and (b) is not caused by the effect of the sputtered Cu atom energy. The difference will be
61
discussed in the section on surface morphology. The uniformity of the electrical resistivity, as
well as the thickness of the copper thin film, has been achieved by using the iron cover. This
behavior could be explained by the copper target erosion depth profile uniformity and through
analysis of the AFM image and XRD patterns of the thin film as will be mentioned later.
Figure 4.9. The resistivity of the film under room temperature as a function of the radial
position for cases (a) and (b), where films were deposited at an RF power of 100 W, a
gyratory speed of 40 rpm, a sputtering gas pressure 1.02 Pa and a processing time of 1.5 h.
4.3.4. Surface Morphology of the Film Deposited by the Gyratory Square-Shaped RF
Magnetized Plasma Sputtering Source
The performance of the gyratory square-shaped RF magnetized plasma sputtering
source for materials processing on the microstructure of deposited films have been analyzed
using AFM and XRD. Figures 4.10 (a)–(b) show three-dimensional AFM images of the
deposited films for cases (a) and (b), respectively. The AFM is used to observe the surface
morphology of the films prepared at an RF power of 100 W, an Ar pressure of 1.0 Pa and a
processing time of 1.5 h. As shown in Figs. 4.10 (a)–(b), the AFM images of the copper thin
films illustrate that their surface morphologies are strongly affected by the use of the iron cover.
It can be clearly observed that the iron cover mounted on the magnets for magnetic shielding
in the magnetized plasma sputtering deposition has a significant impact on the grain size and
surface roughness of the copper thin film, which demonstrates the evidence of the structural
62
deformation. Figure 4.10 (a) shows that the films for case (a) exhibit sharp needle-shaped
grains vertical to the substrate with a mean grain size of approximately 40 nm. However, for
case (b) the film surface shows an improved smooth surface [see Fig. 4.10 (b)] with a reduced
grain size, as well as a mean grain size of around 20 nm. Inspecting Figs. 5.10 (a)–(b), it can
be observed that the surface coverage increases and the isolated nanoparticles coalesce together
to form a more uniform film when using the iron cover in the magnetized plasma sputtering
deposition. In addition, the surface roughness of the Cu film has been observed. The
roughnesses of the Cu films are about 3.73 nm and 2.49 nm for cases (a) and case (b),
respectively. Therefore, the surface properties of the AFM images of the Cu film in Fig. 4.10
are in good agreement with the results of the film thickness and the resistivity profiles as shown
in Figs. 4.8 and 4.9.
63
Figure 4.10 (a) – (b): Three-dimensional AFM micrographs of the film prepared at an RF
power of 100 W, a gyratory speed of 40 rpm, a sputtering gas pressure 1.02 Pa and a
processing time of 1.5 hour for case (a), without iron shielding and for case (b), with iron
shielding.
In order to investigate the performance of the gyratory square-shaped RF magnetized
plasma sputtering source on the structure of the Cu film, X-ray diffraction (XRD) standards
(card no. ICDDPDF#04-0836) were used. Figure 4.11 shows the XRD patterns of Cu films
prepared by the gyratory square-shaped RF magnetized plasma sputtering source for cases (a)
and (b). It is seen that both samples exhibit a face-centered-cubic structure. For (004)Si, the
strong diffraction peak was observed at 69.2° and also at 32.94° with a small magnitude. This
is caused by double diffraction. The diffraction positions are (200), (220), (311), (200), (222),
(400), (311), and (420). The peak position is in good agreement for both cases. Two typical
texture components of (111) and (200) were observed in XRD patterns. The relative intensity
of the (111) peak to the (200) [I(111)/ I(200)] is often defined as the film texture. Considering the
ratio of randomly oriented Cu powder, the I(111)/ I(200) values are 13.76 and 4.08. The full width
at half maximum (FWHM) values of the diffraction peaks were 0.42° and 0.66° for cases (a)
and (b), respectively.
64
Figure 4.11. X-ray diffraction (XRD) patterns of Cu films deposited by the gyratory square-
shaped RF magnetized plasma sputtering source for cases (a) and (b).
4.3.5. Aluminum-Doped Zinc Oxide (AZO) Film Prepared by RF Magnetized Plasma
Sputtering Source with Square-Shaped Rod Magnets
The aluminum-doped zinc oxide (AZO) films were prepared with a square-shaped
magnet setup using a high-density radio frequency (RF) magnetized sputtering plasma source
for uniform target utilization. The glass substrate was used. The glass substrates were cleaned
using ultra sound cleaning technology. Eight neodymium permanent rod magnets of 30 × 5 ×
3 mm, where the connection between N-pole and S-pole magnets is one side of the square. The
two-dimensional magnetic field profiles, the film thickness, the resistivity of the films, atomic
force microscopy (AFM), the spectroscopy, and the x-ray diffraction (XRD) profiles have been
investigated for case (a) stationary and case (b) rotational speed of 45 rpm. The experiments
were performed in a stainless-steel cylindrical vacuum chamber with an outer diameter of 235
mm, an inner diameter of 160 mm and a height of 195 mm. Argon (Ar) gas at 0.5 Pa was used
by regulating a flow meter, an RF power of 120 W at 13.56 MHz was applied to the target via
an impedance matching network. The sputtering processing time was 2.25 h. The film thickness
and resistivity of the prepared AZO films were measured at r = 0 mm for both case (a) and (b).
Prepared aluminum-doped zinc oxide (AZO) films on the glass substrate with a square-shaped
magnet setup for case (a) stationary and case (b) rotational speed of 45 rpm are shown in Fig.
4.12. It is found that the AZO films were not transparent because of high input RF power and
long processing time.
65
(a)
(b)
Figure 4.12. Prepared aluminum-doped zinc oxide (AZO) films on the glass substrate with a
square-shaped magnet setup for case (a) stationary and case (b) rotational speed of 45 rpm.
Figure 4.12 shows the AFM photographs of the AZO films prepared on the glass
substrate with a square-shaped magnet setup for case (a) stationary and case (b) rotational speed
of 45 rpm. The sharp needle-shaped grains vertical to the substrate were observed for case (b).
However, for case (a) the film surface shows an improved smooth surface with a reduced grain
size. Later, the scanning force microscopy (SEM), raman spectroscopy, x-ray photoelectron
spectroscopy (SPS), x-ray diffraction (XRD), optical emission spectrometer (OES), Ultraviolet
Visible Near-Infra red (UV-Vis-NIR) spectroscopy and Hall measurement techniques will be
analyzed for the deposited AZO films.
66
Figure 4.12. AFM photographs of the AZO films prepared on the glass substrate with a
square-shaped magnet setup for case (a) stationary and case (b) rotational speed of 45 rpm.
67
4.4. Conclusion
In order to assess the developed RF magnetized plasma sputtering source with a
gyratory square-shaped magnet arrangement consisting of eight neodymium bar magnets of 30
mm × 5 mm × 3 mm, the thickness and the resistivity profiles and the microstructure of the
copper thin film deposited on the Si wafer substrate were analyzed, using a copper target of
160 mm × 160 mm × 3 mm. Two magnetic arrangements were investigated to discuss the film
properties: case (a) without iron shielding and case (b) with iron shielding of 5 mm × 3 mm ×
1 mm, which is used for magnetic shielding. Copper films were prepared at an RF power of
100 W, an Ar pressure of 1.0 Pa, a processing time of 1.5 h, a gyratory speed of 40 rpm, and a
target– substrate distance of 45 mm. The magnetic shielding material reduces the horizontal
magnetic flux density, |Bx|, at x = 0 from ~1120–824 G and the axial component, |Bz|, from
~824–706 G at x = -15 and 15 mm. The RF discharge voltages, VRF, (peak-to-peak value) were
1080 V and 1160 V, and the absolute values of the self-bias dc voltages, Vsb, of the electrode
were 528 V and 573 V for cases (a) and (b), respectively. The estimated plasma densities were
8.4 × 109 cm-3 and 1.1 × 1010 cm-3 at r = 5 mm and z = 10 mm for cases (a) and (b), respectively.
A typical star-shaped plasma discharge has been observed. The thickness of the deposited
copper thin film for case (b) is around 0.6 m for 0 < r < 30 mm, which decreases at the edge
of the target, whereas the thickness profile fluctuates significantly for case (a). The roughnesses
of the thickness profile are approximately 24.45% and 8.06% for cases (a) and (b),
respectively. The resistivities are 7.89 × 10-8 m and 4.33 × 10-8 m at r = 30 mm for cases
(a) and (b), respectively. The Cu films for case (a) exhibit sharp needle-shaped grains vertical
to the substrate with a mean grain size of ~40 nm, and for case (b) the film surface shows an
improved smooth surface with a reduced mean grain size of ~20 nm. The roughnesses of the
copper thin films are ~3.73 nm and ~2.49 nm for case (a) and case (b), respectively. An XRD
investigation shows that the FWHM values of the diffraction peak were 0.42° and 0.66° for
cases (a) and (b), respectively and both samples exhibit a face-centered-cubic structure.
Therefore, the proposed gyratory square-shaped RF magnetized plasma sputtering source has
an advantage for making the functional films. The prepared AZO films on the glass substrate
for case (a) stationary and case (b) rotational speed of 45 rpm were not transparent because of
high input RF power and long processing time.
68
References
[1] M. A. Lieberman and A. J. Lichtenberg, Principles of plasma discharges and material
processing, 2nd ed. John Wiley & Sons, Inc., New York, 2005.
[2] F. F. Chen and J. P. Chang, Lecture notes on principles of plasma processing, Kluwer
Academic, New York, 2003.
[3] N. Hershkowitz, IEEE Trans. Plasma Sci., 44, 4, 347-363, 2016.
[4] Y. Ohtsu and T. Yanagise, Plasma Sources Sci. Technol., 24, 034005, 2015.
[5] Y. Ohtsu, N. Matsumoto, J. Schulze, and E. Schuengel, Phys. Plasmas, 23, 033510, 2016.
[6] Y. Ohtsu and N. Matsumoto, J. Vac. Sci. Technol. A, 32, 3, 2014.
[7] Y. Ohtsu, Y. Yahata, J. Kagami, Y. Kawashimo, and T. Takeuchi, IEEE Trans. Plasma Sci.,
41, 8, 1856 – 1862, 2013.
[8] P. Chabert and N. Braithwaite, “Physics of radio frequency plasmas,” Cambridge:
Cambridge University Press, 2011.
[9] Y. Ohtsu and H. Urasaki, Plasma Sources Sci. Technol., 19, 4, 045012-1–6, 2010.
[10] Z. Donko, J. Schulze, U. Czarnetzki, A. Derzsi, P. Hartmann, I. Korolov, and E. Schungel,
Plasma Phys. Control. Fusion, 54, 124003, 2012.
[11] Z. Chen, S. Rauf, and K. Collins, J. Appl. Phys., 108, 073301, 2010.
[12] A. Perret, P. Chabert, J. P. Booth, J. Jolly, J. Guillon, and Ph Auvray, Appl. Phys. Lett.,
83, 243, 2003.
[13] Y. Ohtsu and H. Fujita, Appl. Phys. Lett. 85, 4875, 2004
[14] F. Haase, D. Lundin, S. Bornholdt, and H. Kersten, Plasma Phys., 55, 701, 2015.
[15] M. A. Lieberman and V. A. Godyak, IEEE Trans. Plasma Sci., 26, 3, 955–986, 1998.
[16] T. Mussenbrock, R. P. Brinkmann, M. A. Lieberman, A. J. Lichtenberg, and E. Kawamura,
Phys. Rev. Lett., 101, 8, 085004-1–4, 2008.
[17] J. Schulze, Z. Donko, A. Derzsi, I. Korolov, and E. Schuengel, Plasma Sources Sci.
Technol., 24, 1, 2015.
[18] N. K. Bastykova, Z. Donko, S. K. Kodanova, T. S. Ramazanov, and Z. A. Moldabekov,
IEEE Trans. Plasma Sci., 44, 4, 545-548, 2016.
[19] E. Kawamura, M. A. Lieberman, and A. J. Lichtenberg, Phys. Plasma, 13, 5, 053506-1–
053506-14, 2006.
[20] G. Y. Park, S. J. You, and J. K. Lee, Phys. Rev. Lett., 98, 8, 085003-1–4, 2007.
[21] O. A. Popov, and V. A. Godyak, J. Appl. Phys., 57, 1, 53–58, 1985.
69
[22] V. A. Godyak and R. B. Piejak, Phys. Rev. Lett., 65, 8, 996–999, 1990.
[23] Y. Okuno, Y. Ohtsu, C. Komatsu, and H. Fujita, J. Appl. Phys., 73, 4, 1612–1616, 1993.
[24] Y. Ohtsu, K. Hino, T. Misawa, M. Akiyama, and K. Yukimura, Transactions of the
Materials Research Society of Japan, 36, 1, 99-102, 2011.
[25] J. Schulze, Z. Donko, D. Luggenholscher, and U Czarnetzki, Plasma Sources Sci. Technol.,
18, 3, 2009.
[26] X. Z. Jiang, Y. X. Liu, W. Q. Lu, Z. H. Bi, F. Gao, and Y. N. Wang, Vacuum, 86, 881-
884, 2012.
[27] J. Schulze, E. Schungel, and U. Czarnetzki, J. Phys. D: Appl. Phys., 42, 092005, 2009.
[28] A. R. Petre, M. Bazavan, V. Covlea, V. V. Covlea, I. I. Oprea, and H. Andrei, Rom. Rep.
Phys., 56, 2, 271–276, 2004.
[29] M. Sugawara and T. Asami, Surface Coating Technol., 74./75, 1, 355–357, 1995.
[30] H. Barankova and L. Bardos, Surface Coating Technol., 146–147, 486–490, 2001.
[31] L. Bardos, H. Barankova, and Y. A. Lebedev, Surface Coating Technol., 163/164, 654–
658, 2003.
[32] J. Musil, Vacuum, 50, 3/4, 363–372, 1998.
[33] S. Rahmane, M. S. Aida M. A. Djouadi, and N. Barreau, Superlattices and Microstructures,
79, 148-155, 2015.
[34] H. M. R. Giannetta, C. Calaza, D. G. Lamas, L. Fonseca, and L. Fraigi, Thin Solid Films,
589, 730-734, 2015.
[35] B. Xu, X. G. Ren, G. R. Gu, L. L. Lan, and B. J. Wu, Superlattices and Microstructures,
89, 34-42, 2016.
[36] W-G. Wang, M. Li, S. Hageman, and C. L. Chien, Nature Materials, 11, 64-68, 2012.
[37] T. Konishi, K. Yukimura, and K. Takaki, Surface and Coatings Technol., 286, 239–245,
2016.
[38] M. A. Hossain, T. Ide, K. Ikari, and Y. Ohtsu, Vacuum, 128, 219-225, 2016.
[39] H. Qiu, F. Wang, P. Wu, L. Pan, and Y. Tian, Vacuum, 66, 447-452, 2002.
[40] S. Shanmugan and D. Mutharasu, IEEE – ICSE Proc., 978-1-4673-2395-6/12, 132– 136,
2012.
[41] A. Sreedhar, J. H. Kwon, J. Yi, J. S. Kim, and J. S. Gwag, Materials Sci. in Semiconductor.
Processing., 49, 8–14, 2016.
[42] http://laacg.lanl.gov/laacg/services/serv_codes.phtml.
70
Chapter 5. Ring-Shaped Plasma for Target Utilization Obtained with
Circular Magnets Monopole Setups
5.1. Introduction
CCP sources [1-15] driven by radio frequency (RF) power supply at 13.56 MHz are
widely used functional film preparation and material processing due to its simple experimental
arrangement, an easy maintenance as well as large-diameter substrate processing. However, it
is difficult to produce high-density plasma by CCP source. In particular, the CCP has some
limitations such as (1) low plasma density of less than 109 – 1010 cm-3, (2) low deposition rate,
(3) standing wave effect [16] on the powered electrode (target), and (4) difficulty of controlling
external parameters as well as energized ions at the target independently [3]. Thus, the
conventional CCP is not a good tool for plasma processing. A lot of theoretical [1,6,7,11] and
experimental [9, 10, 17-19] analysis have been studied from the viewpoint of the very high
frequency (VHF) operation. Experimentally [17-19], in order to obtain higher densities than
1010 cm-3 the VHF capacitive discharge is used. The other problem of the standing wave effect
[16] on the powered electrode has been reported in the numerical and the experimental studies
[1, 6, 7, 9-11, 17-20], so that there is considerable plasma non- uniformity. To improve the
plasma uniformity, some methods such as the segmented electrode [21], phase-shift control
[22] have been analyzed. However, the problem is not solved completely as far as the operation
of VHF introduction. In general, it is well known that the hollow-cathode discharges [23–32]
can produce a high-density plasma with only a simple structure of the hollow electrode, because
hollow trench can confine electrons effectively.
Recently, the RF magnetized plasma sources have been widely used in the
microelectronics [33–35] such as magnetic films [36], surface treatment and cleaning [30],
diamond-like carbon [37], biomaterial thin films, flat panel display fabrication; transparent
conductive oxides film preparation [33-35] for solar cells and mobile phones and many other
rapidly growing areas [1-2]. A target utilization of 88% rotational square-shaped arrangement
of rod magnets in RF magnetized plasma [38], 60-70% for planar magnetron sputtering using
a rotating tilted unbalanced and asymmetrical yoke magnet [39] and 59% for a magnet spacing
of 10 mm is obtained in the RF magnetron ring-shaped plasmas based on stationary monopole
arrangements [40]. It is also reported that the target erosion is very low near the chamber wall
[38-41]. S. Wickramanayaka et al reported that the development of a large area high-density
71
plasma source with higher radial plasma uniformity for large area wafer processing was
investigated with multipole arrangement of the magnets [42].
In general, plasma does not exist in outer region because of the presence of the ion
sheath near the chamber wall. It is required to produce (1) outer ring-shaped and (2) specific
area plasmas for obtaining (1) uniform, (2) high-density plasma as well as (3) convenient outer
target area erosion profile near the chamber wall. So, the high-density plasma is necessary an
outer region near the chamber wall for specific target area application from the viewpoint of
target utilization. Our idea is to use the mono-pole magnet arrangement in order to overcome
this problem.
The objective of this chapter is to make an RF magnetized outer ring-shaped plasma
sputtering source with a concentrically monopole arrangement of magnets at various different
magnet-gap distances for the specific area target utilization. The discharge characteristics,
electron temperature, plasma density and electron collision frequency in the monopole
arrangements are considered. In Section 6.2, the experimental setup is explained in detail. In
Section 6.3, results and discussions on the effects of monopole schemes, ring-shaped plasma
discharge, and the plasma parameters are described. In Section 6.4, the results obtained in this
chapter are summarized.
5.2. Experimental Methods
The schematic diagram of the experimental setup is shown in Fig. 5.1. A stainless-
steel cylindrical vacuum chamber with an outer diameter of 235 mm, an inner diameter of 160
mm and a height of 195 mm is used to perform experiments. The vacuum chamber was
evacuated to a base pressure of 2.0 10-5 Pa using an oil rotary pumps and a turbo molecular
before starting the experiment. Ar of 1.0 Pa was introduced into the vacuum chamber at a
working gas pressure through a regulating valve. An iron yoke is used to hold magnets as
monopole arrangement. The spacings between the magnets of different circles for monopole
arrangement were chosen by R= 5, 20 and 35 mm for case (a), (b) and (c) as will be mentioned
later, respectively. Figs. 5.2(a)–(c) show cross sections of magnet arrangements under the
target at R = 5, 20 and 35 mm, respectively.
The magnets are holding by iron yoke disk of 140 mm in diameter are put on the target
plate. A cooling fan was used to cool the magnets and the target. An aluminum plate of 140 ×
140 × 3 mm was used as the target to investigate the discharge and the plasma characteristics
as well as to avoid thin film deposition to a glass window at the bottom for the proposed
72
monopole magnet arrangements in the RF magnetized plasma sputtering source. The target is
input by an RF power of 40 W at the driving frequency of 13.56 MHz through an impedance
matching network and a blocking capacitor, while the vacuum chamber wall acts as a grounded
electrode. During experiments, the RF discharge voltage between the RF powered electrode
and the grounded wall and the DC self-biased voltage were measured by a high-voltage probe
with attenuator of high input impedance and digital oscilloscope with a sampling frequency of
2 G samplings/sec. The plasma densities at various positions are obtained by the measurements
of current-voltage characteristics curve. The current-voltage characteristics curve is measured
by a small cylindrical tungsten probe of 1.0 mm in diameter and 10 mm in length.
Figure 5.1. Schematic diagram of the proposed circular ring-shaped RF
magnetized plasma source by monopole magnet arrangement.
73
The biasing voltage was changed from -70 V to +70 V. The Langmuir probe was
compensated with an LC parallel filter circuit (choke) [3,4,13,15,19,23,40] to avoid the
influence of the RF plasma potential fluctuations on the probe current-voltage characteristics.
The probe measurements were carefully performed so as to minimize the disturbance to the
stable plasma. Moreover, the measurement position was fixed at an axial distance, z = 10 mm
from the target surface. In this position, the magnetic field is perpendicular to the probe surface.
So that electrons are easily collected by the Langmuir probe. In fact, as shown in Fig. 5.7,
electron saturation current is not decreased. Moreover, the probe can collect electrons without
an influence of magnetic field.
(a) (b)
74
(c)
Figure 5.2(a)-(c). Proposed circular monopole magnet arrangement for (a) R = 5 mm, (b) R
= 20 mm, and (c) R = 35 mm.
The measurement positions were selected in the radial positions of r = 45, 42 and 38
mm, where high luminous ring-shaped plasmas discharges are observed in the outer region in
the chamber for the three cases (a): R = 5, (b): R =20 and (c): R = 35 mm, respectively. The
plasma density and electron temperature were estimated from the current-voltage
characteristics curve [1-3] and their values at various radial positions were measured by moving
the L-shaped probe to the radial direction as shown in Fig. 5.1. The measurement positions
were selected in the radial positions of r = 45, 42 and 38 mm, where high luminous ring-shaped
plasmas discharge are observed in the outer region in the chamber for the three cases (a): R =
5, (b): R =20 and (c): R = 35 mm, respectively The origins of axial distance z and radial r
positions are defined at the surface and at the center of the RF powered target, respectively as
shown in Fig. 5.1.
75
5.3. Results and Discussions
5.3.1. Magnetic Field Profile Analysis
The proposed circular monopole magnet arrangements for (a) R = 5 mm, (b) R = 20
mm and (c) R = 35 mm are shown in Figs. 5.2(a)–(c), respectively. Here, “R” is the gap distance
between magnets in consecutive circles. The number of magnets positioned at the center, first
circle, second circle and third circle are 1, 5, 12 and 16, respectively for (a) R = 5 mm set up.
For (b) R = 20 mm set up, 1 and 15 magnets are placed at the center and first circle, respectively,
while for (c) R = 35 mm, 1 and 22 magnets are kept at the center and first circle, respectively.
Second circle or more circle is not used for (b) R = 20 mm and (c) R = 35 mm setups, because
the target effectively covers radial area from r = -50 mm to r = + 50 mm. The neodymium
cylindrical magnets with a surface magnetic flux density of 4120 Gauss, diameter of 10 mm,
and height of 5 mm are positioned on circular iron yoke disk 140 mm in diameter. In order to
investigate the effect of magnetic field pattern in the proposed monopole magnet schemes on
plasma production, 2D magnetic flux lines for three magnets arrangements such as (a) R = 5
mm, (b) R = 20 mm and (c) R = 35 mm are as shown in Figs. 5.3(a)–(c), respectively.
The 2D magnetic flux lines and their radial profiles are analyzed by conventional
magnetic-field analysis software poisson superfish [43] developed in Los Alamos National
Laboratory. As seen in Figs. 5.3(a)–(c), the magnetic field lines generated from N-pole of
magnets pass through the target, airspace, iron yoke and then return to the S-pole for all
monopole magnet arrangement. It is found that the axis symmetry property is satisfied for
monopole arrangement because the flux lines profiles for a half region from r = -80 to 0 mm
have a similar profile as for half region r = 0 to 80 mm [see Figs. 5.3(a)–(c)]. It is important
for generating the high-density magnetized discharge plasma at the outer and specific areas to
use effectively the E×B drift motion, where E and B are the electric field perpendicular to the
target and the magnetic flux density parallel to the target, respectively. In especial, the magnetic
flux density component parallel to the target is significant for the E×B drift motion.
76
(a)
(b)
(c)
Figure 5.3(a)-(c). 2D magnetic flux lines distribution in the proposed RF magnetized ring-
shaped plasma near the target surface for (a) R = 5 mm, (b) R = 20 mm and (c) R = 35 mm.
77
Figure 5.4 (a) and (b) shows the radial profile of absolute value of horizontal magnetic
flux density Bx and vertical magnetic flux density Bz at three magnet arrangements of (a) R
= 5 mm (b) R = 20 mm (c) R = 35 mm at the axial distance of z = 5 mm from the target surface,
respectively. It is found that multi-peaks are obtained at all setups. The pattern of their peaks
for the set up (b) is the same as that for the setup (c) and their magnetic flux density profiles
have six peaks. The setup (a) of R = 5 mm has two large peaks of 590 G at a radial position of
r= 52 mm. The second peaks with a weak value of 168 G are also observed at r = 35 mm.
However, other peak amplitudes are less than 100 G at r = 0 to 35 mm. The three highest
peaks are of 504 G, 370 G, and 371 G at a radial position of r= 38, 23 and 7 mm,
respectively for the set up (b). In the set up (c), the three highest peaks of 491 G, 475 G, and
473 G are also observed at a radial position of r= 53, 37 and 8 mm, respectively. Moreover,
the other magnetic flux peaks from r = 0 to 40 mm for setup (a) are the lowest among three
setups. Their high-value peaks will give some ring-shaped plasma discharges due to the E B
drift effect.
78
(a)
(b)
Figure 5.4. Distributions of the absolute value of the (a) horizontal magnetic flux density,
Bx and (b) vertical magnetic flux density Bz for (a) R = 5 mm, (b) R = 20 mm and (c) R =
35 mm at the axial distance of z = 5 mm from the target surface.
79
5.3.2. Discharge Characteristics
Figures 5.5(a)–(c) show typical images of RF magnetized plasma discharge at an RF
power of 40 W, argon gas pressure of 1 Pa for (a) R = 5 mm, (b) R = 20 mm and (c) R = 35
mm. The RF magnetized plasma discharge image is taken by a low-resolution digital camera.
Here, dashed circles denote the insulated wall positions. Multiple ring-shaped plasma
discharges are not seen for the case (a): R = 5 mm, whereas for the cases (b): R = 20 mm, and
(c): R = 35 mm, multiple ring plasma discharges are observed throughout the center to outer
target area [see Fig. 5.5(a)-(c)]. Moreover, drift motion of electrons has an important effect on
the distribution of plasma discharge. This is because charged particles are strongly confined by
the E B drift effect.
(a)
80
(b)
(c)
Figures 5.5. Typical images of RF magnetized plasma discharge at an RF power of 40 W
and an argon gas pressure of 1 Pa for (a) R = 5, (b) 20 and (c) 35 mm.
81
Figures 5.6 (a) and (b) show the RF discharge voltage, VRF (peak to peak) and absolute
value of the DC self-biased voltage, |Vsb| of the cathode as a function of argon gas pressure for
the three proposed monopole magnet arrangements such as (a) R = 5 mm, (b) R = 20 mm, (c)
R = 35 mm. The argon gas pressure flowing into the chamber was changing from 1 to 6 Pa and
the RF power was fixed at 40 W. The RF discharge voltage decreases gradually from
approximately 420 to 340 Vp-p with increasing argon gas pressure for case (a), because electron
collision frequency increases with increasing Ar gas pressure so that ionization rate increases.
It is seen that the RF discharge voltage does not almost depend on Ar gas pressure for case (b).
This is ascribed from the hall parameter effect, which represents the degree of magnetization
[1-3]. However, for case (c) discharge voltage, VRF falls down up to 4 Pa and then saturates
with rising Ar gas pressure. This saturation of discharge voltage might be predicted from
sustaining the magnetized plasma owing to the magnetic confinement. These results tendency
is almost similar to the tendency in the discharge voltage corresponding to all cases.
(a)
82
(b)
Figures 5.6. (a) Discharge voltage and (b) DC self-biased voltage as a function of argon gas
pressure for (a): R = 5, (b): R =20 and (c): R = 35 mm at a constant RF power of 40 W.
5.3.3. Plasma Characteristics
Figure 5.7 shows typical probe current-voltage characteristics for (c) R = 35 mm
measured at axial distance z = 13 mm from target surface, radial position r = 38 mm, where a
high luminous ring-shaped plasma is observed, and RF power of 40 W. Electron temperature
and plasma density are estimated by this probe current-voltage curve. Figure 5.8 (a) shows
electron temperature as a function of argon gas pressure for setup (a) R = 5, (b) 20 and (c) R=
35 mm at a constant RF power of 40 W and z = 13 mm away from the target surface. Here, the
probe current-voltage curves are measured at radial positions of r = 45, 42 and 38 mm where
the high luminous ring-shaped plasma are observed for R = 5, 20 and 35 mm as shown in Figs.
5.5(a)–(c), respectively. It is found that the electron temperature decreases with increasing gas
pressure for all cases. The electron temperatures are 2.42, 1.71 and 1.15 eV at Ar gas pressure
of 4 Pa for the setup (a), (b) and (c), respectively. However, the setup (a) has the highest
electron temperature, whereas the setup (c) has the lowest electron temperature for all argon
gas pressures.
83
Figure 5.7. Typical I-V Langmuir probe characteristics measured at axial distance z = 13 mm
from target surface for (a): R = 5, (b): R =20 and (c): R = 35 mm, and an RF power of 40 W
and argon gas pressure of 3 Pa.
Plasma density as a function of argon gas pressure for (a) R = 5, (b) 20 and (c) 35 mm
at a constant RF power of 40 W and z = 13 mm is shown in Fig. 6.8 (b). It is seen that the
plasma density increases with increasing gas pressure for all setups. The plasma density is
approximately same for set up (b) and (c) at all gas pressure. The highest plasma densities were
6.26×1015, 1.06×1016 and 1.11×1016 m-3 at 5 Pa for setups (a), (b) and (c), respectively. This is
reasonable because electron temperature is dropping with rising gas pressure. The electron
mean free path and electron-neutral collision frequency are estimated from electron
temperature, gas density, and electron neutral cross-section.
84
(a)
(b)
Figures 5.8. (a) Electron temperature and (b) plasma density as a function of argon gas
pressure for (a): R = 5, (b): R =20 and (c): R = 35 mm at a constant RF power of 40 W and z
= 13 mm away from the target surface.
85
Figures 5.9 (a) and (b) depict the electron mean free path and electron-neutral collision
frequency as a function of argon gas pressure for R = 5, 20, 35 mm at a constant RF power of
40 W and z = 13 mm away from the target surface, respectively. It is seen that the electron
mean free path decreases for Par ≤ 3.0 Pa and then increases with increasing gas pressure for
Par ≤ 5.0 Pa for setups (a), and (b). However, for setup (c), the electron mean free path
decreases abruptly for Par ≤ 2.0 Pa then increases with increasing gas pressure for Par ≤ 5.0
Pa. It is found that the electron mean free path is 41.4, 63.17 and 84.66 mm at Ar gas pressure
of 5 Pa for the setup (a), (b) and (c), respectively.
It is found that the electron neutral collision frequency increases for Par ≤ 2.0 Pa and
then decreases slowly with increasing gas pressure from 3.0 to 5.0 Pa. for set up (b) and (c).
However, for case (a) R = 5 mm set up, the electron collision frequency increases abruptly for
Par ≤ 2.0 Pa then remain constant for Par ≤ 4.0 Pa as well as decreases sharply for Par ≤5.0
Pa. These phenomena are caused by Ramsauer effect [1]. Electron neutral collision frequency
contributes to the generation of plasma charged particles. Electron neutral collision frequency
for case (a) R = 5 mm set up is higher than case (b) R = 20 mm and case (c) R = 35 mm at a
constant RF power of 40 W and z = 13 mm axial distance from the target surface. This means
the formation of outer ring-shaped plasma near the chamber wall for case (a) R = 5 mm is better
than for case (b) R = 20 mm and case (c) R = 35 mm set up. However, outer ring-shaped plasma
for case (b) and case (c) also produced in specific regions. Ion mean free path as a function of
argon gas pressure for R = 5, 20, 35 mm at a constant RF power of 40 W is also investigated.
It is also found that the ion mean free path is independent of argon gas pressure for all cases
with decreasing tendency.
86
(a)
(b)
Figures 5.9. (a) Electron mean free path and (b) electron-neutral collision frequency as a
function of argon gas pressure for (a): R = 5, (b): R =20 and (c): R = 35 mm at a constant RF
power of 40 W and z = 13 mm away from the target surface.
87
5.3.4. Ion Saturation Currents Profile
In the previous work [42], optimization of plasma density and radial uniformity using
multiple arrangements of magnets, there was a marked difference in the normalized ion density
measured at z = 10 mm and z = 58 mm, which decrease sharply during the period 100 to 200
mm. Moreover, the normalized ion density measured at z = 10 mm has more fluctuations with
a lower value than z = 58 mm data. Overall, the ion density radial profile declines abruptly for
all configurations throughout from the center to an outer wall of the chamber [42]. However,
in our works radial profile of ion saturation current has been improved by the monopole
arrangement [see Fig. 5.10]. Figure 5.10 shows that radial profiles of ion saturation current for
the three cases (a): R = 5, (b): R =20 and (c): R = 35 mm at argon gas pressure of 1.0 Pa, a
constant RF power of 40 W by Langmuir probe biased at Vp = -68V and axial distance z = 10
mm away from the target surface. According to probe theory, the ion saturation current is
proportional to the ion flux to the target. It is shown that for the case (b): R =20 mm, the ion
saturation current is almost constant of 50×10-4 A for 0 < r < 45 mm. In contrast, in the case of
case (c): R =35 mm, the ion saturation current has a maximum of 66×10-4 A at r = 5 mm with
keeping almost constant value for 0 < r < 35 mm then decreases gradually with radial position.
On the other hand, for case (a): R = 5 mm, the ion saturation current is the smallest among the
other cases, however, increases significantly for 30 < r < 48 mm. That is, the case (b) is the
best monopole arrangement of magnets from the viewpoint of the target utilization from center
to outer regions.
88
Figure 5.10. Ion saturation current as a function of radial distance for the three cases (a): R
= 5, (b): R =20 and (c): R = 35 mm at argon gas pressure of 1.0 Pa, a constant RF power of
40 W by a Langmuir probe biased at Vp = 68V and axial distance z = 10 mm away from
the target surface.
89
5.4. Conclusions
The outer ring-shaped RF magnetized plasma has been proposed by three circular
monopole magnet arrangement such for (a) R = 5 mm, (b) R = 20 mm and (c) R = 35 mm. It is
found that the axis symmetry property in monopole arrangement because the flux lines profiles
for half region from r = -80 to 0 mm have a similar profile as for a half region r = 0 to 80 mm.
It is seen that a high luminous ring-shaped plasma is observed in contiguity with the wall
position of r = 47-50 mm for (a) R = 5 mm, whereas multi-ring discharges are observed for (b)
R = 20 and (c) 35 mm. The tendency of the absolute value of the DC self-biased voltage |Vsb|
of the cathode is almost similar to the pattern in the discharge voltage |Vrf| corresponding to the
three monopole schemes. It is found that the electron temperature decreases with increasing
gas pressure for all cases. However, it is seen that the plasma density increases with increasing
gas pressure for all setups. The smaller gap distance, the fluctuations of the plasma density and
ion saturation current are more. However, a big gap distance as (c): R = 35 mm, has high ion
saturation current with more fluctuations comparison with case (b): R = 20 mm. So that case
(b): R = 20 mm is effective to improve target utilization from center to the outer area. From a
viewpoint of the center to outer target utilization, case (b): R =20 mm is more effective than
other two cases. However, case (a): R = 5 mm can be used to improve target utilization only in
the outer area, whereas (c): R = 35 mm has the highest ion saturation current and lowest electron
temperature among all cases.
90
References
[1] M. A. Lieberman and A. J. Lichtenberg, Principles of Plasma Discharges and Material
Processing, 2nd ed. John Wiley & Sons, Inc., New York, 2005.
[2] F. F. Chen and J. P. Chang, Lecture Notes on Principles of Plasma Processing, Kluwer
Academic, New York 2003.
[3] Y. Ohtsu and T. Yanagise, Plasma Sources Sci. Technol., 24, 034005, 2015.
[4] Y. Ohtsu, N. Matsumoto, J. Schulze, and E. Schuengel, Phys. Plasmas, 23, 033510, 2016.
[5] B. P. Wood, M. A. Lieberman, and A. J. Lichtenberg, IEEE Trans. Plasma Sci., 23, 89,
1995.
[6] S. Rauf and M. J. Kushner, IEEE Trans. Plasma Sci., 27, 1329, 1999.
[7] P. Chabert, J.L. Raimbault, J. M. Rax, and M. A. Lieberman, Phys. Plasma 11, 1775, 2004.
[8] T. Mussenbrock, D. Ziegler, and R. P. Brinkmann, Phys. Plasma, 13, 083501, 2006.
[9] Z. Donko, J. Schulze, U. Czarnetzki, A. Derzsi, P. Hartmann, I. Korolov, and E. Schungel,
Plasma Phys. Control. Fusion, 54, 124003, 2012.
[10] Y-X. Liu, F. Gao, J. Liu, and Y-N. Wang, J. Appl. Phys., 116, 043303, 2014.
[11] S. Sharma and M. M. Turner, J. Phys. D: Appl. Phys., 47, 285201, 2014.
[12] E. Abdel-Fattah, M. Bazavan, and H. Sugai, Phys. Plasmas, 19, 113503, 2012.
[13] J. P. Booth, G. Curley, D. Maric, and P. Chabert, Plasma Sources Sci. Technol., 19,
015005, 2010.
[14] Y. Ohtsu and H. Fujita, Appl. Phys. Lett. 92, 171501, 2008.
[15] I. Lee, D. B. Graves, and M. A. Lieberman, Plasma Sources Sci. Technol., 17, 015018,
2008.
[16] Y. Ohtsu, S. Tsuruta, T. Tabaru, and M. Akiyama, Surface Coatings Technol., 307, 1134,
2016.
[17] P. Chabert and N. Braithwaite 2011 Physics of Radio- Frequency Plasmas (Cambridge:
Cambridge University Press).
[18] Y. Yang and M. J. Kushner, J. Appl. Phys., 108, 113306, 2010.
[19] V. A. Godyak and O. Popov, Sov. J. Plasma Phys., 5, 227, 1979.
[20] Y. Ohtsu and H. Fujita J. Appl. Phys., 73, 2155, 1993.
[21] A. Perret, P. Chabert, J-P. Booth, J. Jolly, J. Guillon, and Ph. Auvray, Appl. Phys. Lett.
83, 243, 2003.
[22] Y. Yang and M. J. Kushner, J. Appl. Phys., 108, 113306, 2010.
[23] V. Volynets, H. Shin, D. Kang, and D. Sung D, J. Phys. D: Appl. Phys., 43, 085203, 2010.
91
[24] Y. Ohtsu and H. Urasaki, Plasma Sources Sci. Technol. 19, 045012, 2010.
[25] S. Hashiguchi and M. Hasikuni, Jpn. J. Appl. Phys., 26, 271, 1987.
[26] T. Lafleur and R. W. Boswell, Phys. Plasma, 19, 023508-1–12, 2012.
[27] H. Pak and M. J. Kushner, J. Appl. Phys. 71, 94, 1992.
[28] H. Eichhorn, K. H. Schoenbach, and T. Tessnow, Appl. Phys. Lett. 63, 2481, 1993.
[29] A. R. Petre, M. Bazavan, V. Covlea, V. V. Covlea, I. I. Oprea, and H. Andrei, Rom. Rep.
Phys., 56, 271, 2004.
[30] M. Sugawara and T. Asami, Surface Coating Technol., 74/75, 355, 1995.
[31] H. Barankova and L. Bardos, Surface Coating Technol., 146–147, 486, 2001.
[32] L. Bardos, H. Barankova, and Y. A. Lebedev, Surface Coating Technol. 163/164, 654,
2003.
[33] J. Musil, Vacuum, 50, 363, 1998.
[34] S. Rahmane, M. S. Aida M. A. Djouadi, and N. Barreau, Superlattices and Microstructures,
79, 148, 2015.
[35] H. M. R. Giannetta, C. Calaza, D. G. Lamas, L. Fonseca, and L. Fraigi, Thin Solid Films,
589, 730, 2015.
[36] B. Xu, X. G. Ren, G. R. Gu, L. L. Lan, and B. J. Wu, Superlattices and Microstructures,
89, 34, 2016.
[37] W-G. Wang, M. Li, S. Hageman, and C. L. Chien, Nature Materials, 11, 64, 2012.
[38] T. Konishi, K. Yukimura, and K. Takaki, Surf Coatings Technol. 286, 239, 2016.
[39] M. A. Hossain, T. Ide, K. Ikari, and Y. Ohtsu, Vacuum, 128, 219, 2016.
[40] T. Iseki, Vacuum, 84, 339, 2010.
[41] Y. Ohtsu, M. Shigyo, M. Akiyama, T. Tabaru, Vacuum, 101, 403, 2014.
[41] S. Wickramanayaka, Y. Nakagawa, Y. Sago, and Y. Numasawa, J. Vac. Sci. Technol. A:
18, 823, 2000.
[42] http://laacg.lanl.gov/laacg/services/serv_codes.phtml.
92
Chapter 6. Outer Ring-Shaped Radio Frequency Magnetized Plasma
Source for Target Utilization in Specific Area
6.1 Introduction
The development and application of magnetron sputtering systems have found
widespread use in many industrial manufacturing sectors for physical vapor deposition
[1–2]. Magnetron sputtering techniques use mutually perpendicular magnetic and electric
fields, which confine the plasma close to the deposition source, so increasing the
probability of gas ionization and therefore the plasma density and sputter rates [3–4]. Due
to an ion bombardment, the deposition species usually in metals, alloys, ceramics and
polymers are atomized significantly by sputtering from the target (cathode). On the other
hand, CCP discharges functioned by radio frequency (RF) power supply are widely used
for a great variety of technological applications ranging from plasma dry etching [5] and
medical applications such as sterilization or wound healing [6–7], surface treatment and
cleaning [8], microelectronics [9–11] such as magnetic films [12], diamond like carbon
[13], biomaterial thin films, flat panel display fabrication; transparent conductive oxides
film preparation for solar cells [9], [11], [14], mobile phones and many other rapidly
growing areas in material processing [3–4], [7]. Because the CCP discharges are in
principle of the simplest setup and easily maintenance capable techniques for large-
diameter substrate processing in microelectronic device fabrication, work in various
gases [15–24].
Generally, near the chamber wall, the plasma potential and the plasma density is
very small because an ion sheath exists near the chamber wall [3], [24]. The plasma
potential varies slowly in the plasma, but rapidly in the sheath region. Only in the sheath
region, the quasineutrality property cannot be satisfied. The plasma density profile is
relatively flat in the center and falls sharply near the sheath edge [4], [24]. Moreover, in
practical industrial application, the outer width of the target is large. Therefore, it is
required a high-density plasma discharge in a specific area and also in the outer region of
the chamber to obtain convenient outer target area erosion profile near the chamber wall.
Moreover, to deposit a functional thin film in a specific area and near the chamber wall,
the target utilization in a specific area is required.
The main idea is to use a CCP discharges by the mono-pole magnet arrangement
at a fixed magnet-gap distance in order to get high density plasma in the outer region. In
our previous work [25], we have developed RF ring-shaped plasma obtained with circular
93
magnet monopole arrangement. In the conventional magnetron system, target material is
not effectively used because the high-density plasma is localized on the target surface.
The target utilization is very low approximately at 20-30%. From the practical viewpoint
of the limited resources, the utilizations of the target material are necessary. The
symmetrical magnets magnetron sputtering method with one inner magnet and two outer
annular magnets facing each other was investigated [26]. The maximum target erosion
rate was 57%. The rotating magnet sputtering has also been proposed by rotating helical
magnets for increasing the target utilization efficiency [27]. Using a rotating unbalanced
and asymmetrical magnet, a flat erosion-sputtering method has been developed [28]. The
estimated target utilization had a value of 80% and 77% for 5-in, 4-in aluminum target
material, respectively. The target utilization efficiency was increased from 73.6% to
86.3% when iron pole pieces were used in the rotating cruciform arrangement of
neodymium magnets. However, effect of outer ring-shaped plasma discharge near the
chamber wall in a specific area had not been investigated in detail.
The objective of this work is to produce a outer ring-shaped and specific area
plasmas for obtaining high-density plasma and outer target erosion profile near the
chamber wall. In Section 6.2, the proposed experimental setup of CCP discharges is
explained in detail. In Section 6.3, the results and discussions on the effects of the three
setups are investigated in a monopole circular magnet arrangement of the magnetic flux
lines and their profiles, discharge characteristics, plasma discharge luminescence, plasma
density, and ion saturation currents are explained. In Section 6.4, the results obtained in
this work are summarized.
94
6.2 Experimental Setup Details
The walls of the vacuum chamber are an outer diameter of 235 mm, inner
diameter of 160 mm and a height of 195 mm, which was used to carry out the experiments.
The vacuum chamber was evacuated by a combination of an oil rotary pump and a turbo
molecular pump to a very low pressure. Before starting the experiment, the base pressure
in the chamber of 1.6 10-5 Pa is attained. Ar gas was fed into the chamber through a
regulating flow meter at an operating gas pressure of 1.5 Pa. The schematic illustrations
accompanying diagnostics of the proposed outer ring-shaped CCP source for target
utilization in specific areas using magnets monopole arrangement is shown in Fig. 6.1.
The neodymium cylindrical magnets with surface magnetic flux density of 4120 Gauss,
diameter of 10 mm, and height of 5 mm are used. The gap distance between magnets was
fixed at 5 mm in consecutive circles shown in Fig. 6.1. We selected a fixed gap distance
of 5 mm because high luminescent ring-shaped plasma discharge obtained by circular
magnet monopole arrangement [25].
In especial, R = 5 mm is chosen due to the ion saturation current increases
significantly for 30 < r < 49 mm. The number of magnets positioned in setups (a): 1, 5,
11, and 14, (b): 1, 5, and 11, and (c): 1, and 5. An iron yoke was used to hold the magnets
and positioned on the target surface. Fig. 6.2 shows the cross sections of the magnet
monopole arrangements of fixed gap distance of 5 mm, including a center magnet, with
magnets in setups (a): three circles, (b): two circles, and (c): one circle, respectively. To
cool the magnets and target, a cooling fan was used. An RF power of 50 W at 13.56 MHz
was given as input power to the target, where the chamber wall was acted as a grounded
electrode. An impedance-matching network and a blocking capacitor were also used to
connect RF power supply with the target. To investigate the discharge and the plasma
characteristics, an aluminum target of 140 mm × 140 mm × 3 mm was used, in order to
avoid film deposition on the glass window at the bottom in the chamber.
A high-voltage probe with an attenuator of high input impedance and a digital
oscilloscope with a sampling frequency of 2 giga samplings/sec were used to measure the
RF discharge voltage between the RF powered electrode (target) and the grounded
vacuum wall, as well as a DC self-biased voltage. The sputtering rate is directly
proportional to the ion flux incident to the target. The ion saturation currents, proportional
to the ion flux to the target were measured by a negatively biased cylindrical tungsten
probe of 1.0 mm in diameter and 10 mm in length [4]. The measurements of the ion
95
saturation currents and current-voltage (I-V) characteristics were done by moving the L-
shaped probe to the radial direction, as shown in Fig. 1, and the position were set at 0 < r
< 49 mm. The biasing voltage was changed from –65 V to +65 V. The probe wire was
compensated with an LC filter circuit (choke) in order to reduce the influence of the RF
plasma potential fluctuations on the probe current-voltage characteristics [3], [7]. The
probe measurements were carefully performed so as to minimize the disturbance to the
stable plasma and also taking care of the fluctuation of the plasma potential at the
excitation frequency and its higher harmonics. Moreover, the measurement position was
fixed at axial distance, z = 9 mm from the target surface. The magnetic field is
perpendicular to the probe surface in this position, so that electrons are easily collected
by the Langmuir probe. A Cartesian coordinate system is used. As shown in Fig. 6.1, the
origins of the axial distance z and the radial positions r are defined at the surface and at
the center of the RF powered target, respectively.
Figure 6.1. A schematic illustrations of the proposed outer ring-shaped CCP source
for target utilization in specific area using magnets monopole arrangement.
96
Figure 6.2. Cross sections of the magnet monopole arrangements, with a center
magnet, and magnets in setups (a): three circles, (b): two circles, and (c): one circle.
6.3. Results and Discussions
6.3.1. Profiles of Magnetic Field of the Monopole Setups
The absolute value of the horizontal, Bx magnetic flux densities distributions
were analyzed by Poisson Superfish software developed in Los Alamos National
Laboratory. Fig. 6.3 demonstrates magnetic flux density profile in the proposed outer
ring-shaped CCP source for target utilization in specific areas at the axial position, z = 7
mm from the target surface. It is seen that, Bx has only two peaks and the patterns of
their peaks are almost same for setups (b) and (c). The three highest peaks of Bx are (a):
356 G, (b): 486 G, (c): 495 G, at a radial position of r= 49, 37, and 22 mm,
97
respectively. The radial profiles of magnetic flux in additional two cases of 1,11 (center
and 2nd circle: setup (d)) magnets, and 1, 14 (center and 3rd circle: setup (e)) magnets
were also investigated. To discuss the role of the inside magnets for the production of the
plasma, 1 and 11 magnets were placed at the center and in the second circle, 1 and 14
magnets were placed at the center and in the third circle, respectively. Multiple peaks
were obtained in setups (d) and (e). The three highest peaks had values of 386, 222, and
245 G at radial positions of r = ±39, ±23, and ±7 mm, respectively, for setup (d). In the
setup (e), the three highest peaks with amplitudes of 226, 355, and 342 G were also
observed at radial positions of r = ±52, ±36, and ±8 mm, respectively. Moreover, the
other magnetic flux peaks from r = 0 to ±40 mm for setup (a) and r = 0 to ±35 mm for
setup (b) had the lowest amplitudes among the five setups. Moreover, high-amplitude
peaks are observed in the outer area for setups (a) and (b). To discuss the effect of the
outer ring-shape structure on axial profile of magnetic flux at some radial positions, two
dimensional magnetic field lines distributions for setups (a), (b), and (c) were investigated.
The magnetic channeling increases with increasing the magnets in the inside circle.
Moreover, magnetic flux distributions of high-amplitude are restricted at the portion of
the outer area so that inside magnet arrangements for setup (a) and (b) produce high
luminescence plasma near the chamber wall [see Fig. 6.3 and 6.4].
High luminescent ring-shaped plasma discharges were produced by high-value of
Bx due to the E B drift effect at their corresponding position. The magnetic flux density,
Bx for setup (a) have the lowest among three setups from r = 0 to 50 mm. The larmor
radius is known as gyro radius. The gyro radius of electron, rLe will be smaller in
accordance with, , and the cyclotron frequencies, increase in a rising
magnetic field by according to the equation, , where e, m, and ve are electronic
charge, masses of charged particles, and velocity of electron is the speed perpendicular
to the magnetic field, Bx. The gyro radius and cyclotron frequency are calculated with
an assumption of Te = 2 eV at the highest magnetic flux position. It was found that gyro
radius of the electron, re are (a): 0.1519 mm, (b): 0.1111 mm, and (c): 0.1093 mm at r=
49, 37, and 22 mm, respectively. Moreover, the cyclotron angular frequencies of
the electron, were (a): 6.25×109 Hz, (b): 8.54×109 Hz, (c): 8.68×109 Hz, at r= 49,
37, and 22 mm, respectively. It is important to investigate the electron motion in the
magnetic field because the electrons produce plasma. In order to analyze the
magnetization degree of the charged particle, the electron hall parameter was calculated
98
from c/νm, where c and νm are the charged particle cyclotron frequency and the charged
particle-neutral collision frequency, respectively. The neutral collision frequency of
electrons was calculated from the relation of nArve, where , nAr, and ve are the electron
neutral collision cross section, the Argon gas density and the electron velocity,
respectively. The electron hall parameter was 245, 395, and 393 at high luminescent
plasma discharge position for setups (a), (b), and (c), respectively. Thus, electrons are
strongly magnetized in the high luminescent plasma discharge position, i.e. near the
chamber wall.
Figure 6.3. Absolute value of the horizontal magnetic flux densities, Bx in the outer
ring-shaped CCP source with a center magnet, and magnets in setups (a) three circles,
(b) two circles, (c) one circle, (d): center and 2nd circle, and (e): center and 3rd circle
magnets at z = 7 mm from the target surface.
6.3.2. Emission of Outer Ring–Shaped Plasma
Fig. 6.4(a)–(c) show a typical RF magnetized plasma discharge at an RF power
of 50 W, Ar gas pressure of 1.5 Pa, with a center magnet, and magnets in setups (a): three
circles, (b): two circles, and (c): one circle. Ring-shaped plasma in the specific outer area
is observed at the position with the peak magnetic flux density and its diameter depends
on a number of magnet circles. This is because, electrons strongly are affected based on
the E×B drift motion, where the electric field, E, and the magnetic flux density, B, are
99
perpendicular to the target and parallel to the target, respectively. At the high magnetic
flux density position, which produce high luminescent plasma discharge and E×B drift
motion considered to play a significant role in the magnetized plasma discharge. However,
high luminescent plasma discharge was observed in the outer area.
(a)
(b)
100
(c)
Figure 6.4. Typical RF outer ring-shaped RF magnetized plasma discharge at an RF
power of 50 W, Argon gas pressure of 1.5 Pa with a center magnet, and magnets in
setups (a): three circles, (b): two circles, and (c): one circle.
6.3.3. Outer Ring–Shaped Plasma Characteristics
Fig. 6.5 (a) and (b) show the discharge voltage, VRF, and the absolute value of the
DC self-biased voltage, |Vsb|, respectively, in the outer ring-shaped RF magnetized plasma
sputtering source as a function of the Ar gas pressure at an RF power of 50 W, for setups
(a), (b), and (c). The discharge voltage is defined as the peak to peak value of RF voltage
when RF plasma is sustained at some condition. Discharge voltage denotes a voltage level
at which the charged particle collision rate with neutral particles occurred and ionization
of neutrals sustains the plasma in the steady state. The RF discharge voltages, VRF, were
460, 340, and 440 Vp-p and the absolute value self-bias dc voltages, |Vsb|, were 154, 83,
and 143 V at Ar gas pressures of 1.5 Pa, for setups (a), (b), and (c), respectively, at an RF
power of 50 W.
The discharge voltage, VRF, and the self-bias dc voltage, |VSB|, for (b) is the lowest
among the other setups, where setups (a) and (b) have approximately the same value. This
means RF discharge for setup (b) is easier than the other setups because of lower
discharge and self-biased voltage among three cases. In general, the discharge voltage,
VRF, decreases with increasing Ar gas pressure, this is because the electron collision
101
frequency and the ionization rate, increases with an increasing Ar gas pressure, and the
discharge voltage is approximately inversely proportional to the electron collision
frequency at a fixed absorbed power [1–2]. However, in this proposed outer ring-shaped
RF magnetized plasma, the discharge voltage, VRF, is almost constant due to stable plasma
in all Ar gas pressure at an RF power of 50W.
Fig. 6.6 shows typical probe current-voltage characteristics for setups (a), (b), and
(c), respectively. The measurements position was at axial distance z = 10 mm from target
surface, radial position (a): at r = 45 mm, (b): at r = 40 mm, and (c): at r = 33 mm, where
a high glowing ring-shaped plasma is observed, and an RF power of 50 W.
(a)
102
(b)
Figure 6.5. (a) RF discharge voltage, VRF (peak to peak) and (b) absolute value of the
DC self-biased voltage, |VSB| of the target as a function of Argon gas pressure in the
outer ring-shaped RF magnetized plasma at an RF power of 50 W for all setups.
Figure 6.6. A typical current-voltage probe characteristics for monopole arrangements
with a center magnet, and magnets in setups (a): three circles, (b): two circles, and (c):
one circle.
103
6.3.4. Ion Saturation Currents in Outer Ring–Shaped Plasma
In general, the electrons have much smaller gyro radius than ions because ions
mass is more that electrons. In a relatively weak magnetic field, electron gyro radius can
be quite large. The ion current collection will not be affected by stronger fields are
existing. All the electron Hall parameters are much larger than one. Thus, it is suggested
that the electrons are intensely magnetized in the high luminescent plasma discharge
position, i.e. near the chamber wall. In contrast, all the ion Hall parameters are much less
than one, so that, ions are not almost magnetized. In particular, the strong magnetic fields
make the electron gyro radius smaller than the probe radius, which will lower the
magnitude of electron saturation currents, Ies. This is because the probe depletes the field
lines that it intercepts, and further electrons can be collected only if they diffuse across
the magnetic fields. Moreover, the probe cannot collect electron when the magnetic fields
are parallel to the probe.
Fig. 6.7 shows the ion saturation currents, Iisat as a function of the radial distance
from the center of the target in the proposed outer ring-shaped CCP source for obtaining
specific area target erosion profile near the chamber wall by magnetic monopole
arrangements for setups (a), (b), and (c), respectively. The ion saturation currents, Iisat
were measured by an L-shaped Langmuir probe, negatively biased at Vp = -65 V and
positioned in the at an axial distance z = 9 mm away from the target surface. The Iisat are
0.6 mA, 0.79 mA, and 0.46 mA, for setups (a), (b), and (c), respectively at a radial
position of r = 47 mm. The Iisat for setup (a) is the lowest for 0 < r < 35 mm, whereas for
setup (c), Iisat has the highest value among three setups. It is seen that, Iisat is very high in
the outer target region near the chamber wall for setups (a) and (b), where Iisat for setup
(c) decreases slowly. Comparison of the positions of the high luminescent plasma
discharge, peak ion saturation current and peak of the horizontal, Bx magnetic flux
position at the various setups are listed in table 6.1.
104
Figure 6.7. Radial profiles of ion saturation currents in the proposed outer ring-shaped
RF plasma using magnetic monopole arrangements, with a center magnet, and magnets
in setups (a) three circles, (b) with two circles, and (c) one circle.
It is defined that the effective radial width (ra, rb, and rc) as a distance between
the positions, where a level of 0.707 of the maximum value of ion the saturation currents,
Iisat and maximum of Iisat. It was found that ra = 10 mm (39 to 49 mm), rb = 8 mm (37 to
45 mm), and rc = 10 mm (25 to 35 mm). Moreover, the effective radial width position is
shifted from the chamber wall to center with removing the magnets in the outer circles.
The roughnesses of the ion saturation current for setups (a), (b), and (c) were also
calculated using, (Iisat, max - Iisat, min)/(Iisat, max + Iisat, min). The roughnesses of Iisat are 51.3%,
42.6%, and 20.3% for setups (a), (b), and (c), respectively. The ion saturation current is
directly proportional to the ion flux incidence to the target. The measurement of Iisat is
the simplest and the best technique to estimate the plasma density, n. Plasma density is
estimated from ion saturation current using Bohm sheath criteria [3-4]. The calculated
plasma densities were 6.78×1016 m-3, 8.98×1016 m-3, and 5.19×1016 m-3 in the proposed
outer ring plasma for setups (a), (b), and (c), respectively at a radial position of r = 47
mm and axial position of z =9 mm from the target surface.
105
Table 6.1
Comparison of the various positions
Monopole magnet
setups [mm]
Setup (a) Setup (b) Setup (c)
High luminescence plasma
position [mm]
48-50 42-46 22-35
Peak ion saturation current
position [mm]
49 45 35
Peak horizontal magnetic
flux, Bx position [mm]
50 43 23
6.4. Conclusion
Ring-shaped plasma in the specific outer area was observed at the position of the
peak magnetic flux density and its diameter depends on a number of magnet circles. The
high-density plasma was found in the outer target region near the chamber wall for all
type arrangements. The plasma discharge results that the target utilization can be
controlled in the outer specific area near the wall. The electron hall parameter was (a):
245, (b): 395, and (c): 393 at high luminescent plasma discharge position. It was observed
that the ion saturation currents, Iisat are 0.6 mA, 0.79 mA, and 0.46 mA, for setups (a),
(b), and (c), respectively at a radial position of r = 47 mm. The Iisat for setup (a) is the
lowest for 0 < r < 35 mm, whereas for setup (c), Iisat has the highest value among three
setups. It is seen that, Iisat is very high in the outer target region near the chamber wall for
setups (a) and (b), where Iisat for setup (c) decreases slowly. Thus, the proposed CCP
plasma source has an advantage of the outer target erosion profile and target utilization
in the specific area.
106
References
[1] G. Brauer, B. Szyszka, M. Vergohl, and R. Bandorf, Vacuum, 84, 1354-1359, 2010.
[2] N. Hershkowitz, IEEE Trans. Plasma Sci., 44, 4, 347-363, 2016.
[3] M. A. Lieberman, A. J. Lichtenberg, “Principles of plasma discharges and material
processing”, 2nd edn. John Wiley & Sons, Inc., New York, 2005.
[4] F. F. Chen, and J. P. Chang, “Lecture notes on principles of plasma processing”,
Kluwer Academic, New York, 2003.
[5] R. Pillarisetty, “Academic and industry research progress in germanium nanodevices,”
Nature, 479, 7373, 324–328, 2011.
[6] T. Makabe and Z. Petrovic, “Plasma electronics: applications in microelectronic
device fabrication”, London, Taylor and Francis, 2006.
[7] P. Chabert, N. Braithwaite, “Physics of radio frequency plasmas,” Cambridge
University Press, Cambridge, 2011.
[8] H. Barankova and L. Bardos, Surface Coating Technol., 146/147, 486–490, 2001.
[9] S. Rahmane, M. S. Aida M. A. Djouadi, and N. Barreau, Superlattices and
Microstructures, 79, 148-155, 2015.
[10] H. M. R. Giannetta, C. Calaza, D. G. Lamas, L. Fonseca, and L. Fraigi, Thin Solid
Films, 589, 730-734, 2015.
[11] B. Xu, X. G. Ren, G. R. Gu, L. L. Lan, and B. J. Wu, Superlattices and
Microstructures, 89, 34-42, 2016.
[12] W-G. Wang, M. Li, S. Hageman, and C. L. Chien, Nature Materials, 11, 64-68, 2012.
[13] T. Konishi, K. Yukimura, and K. Takaki, Surface Coating Technol. 286, 239–245,
2016.
[14] Tooru Tanaka, T. Terasawa, Y. Okano, S. Tsutsumi, K. Saito, Q. Guo, M. Nishio,
K. M. Yu, and W. Walukiewicz, Solar Energy Materials and Solar Cells, 169, 1-7, 2017.
[15] Takashi Sumiyama, Takaya Fukumoto, Yasunori Ohtsu, and Tatsuo Tabaru, AIP
Advances, 7, 055310-8, 2017.
[16] Y. Ohtsu, N. Matsumoto, J. Schulze, and E. Schuengel, Phys. Plasmas, 23, 033510,
2016.
[17] Y. Ohtsu and N. Matsumoto, J. Vac. Sci. Technol. A, 32, 3, 2014.
[18] Y. Ohtsu, Y. Yahata, J. Kagami, Y. Kawashimo, and T. Takeuchi, IEEE Trans.
Plasma Sci., 41, 8, 1856–1862, 2013.
[19] M. A. Hossain, T. Ide, K. Ikari, and Y. Ohtsu, Vacuum, 128, 219-225, 2016.
107
[20] M. A. Hossain, Y. Ohtsu, and T. Tabaru, Plasma Chem. Plasma Process, 37, 1663–
1677, 2017.
[21] J. Schulze, E. Schungel, and U. Czarnetzki, J. Phys. D: Appl. Phys. 42, 092005, 2009.
[22] J. Schulze, Z. Donko, A. Derzsi, I. Korolov, and E. Schuengel, Plasma Sources Sci.
Technol., 24, 1, 2015.
[23] P. Poolcharuansin, P. Laokul, N. Pasaja, Chingsungnoen A., M. Horprathum, P.
Chindaudom, and J. W. Bradley, Vacuum, 141, 41-48, 2017.
[24] S. Wickramanayaka, Y. Nakagawa, Y. Sago, and Y. Numasawa, J. Vac. Sci. Technol.
A, 18, 3, 823-828, 2000.
[25] Md. A. Hossain and Y. Ohtsu, Jpn. J. Appl. Phys., 57, 01AA05, 2018.
[26] Q. H. Fan, and H. Y. Chen, Thin Solid Films, 229, 143-145, 1993.
[27] J. Musil, J Vac. Sci. Technol. A, 17, 555, 1999.
[28] T. Iseki, Vacuum, 84, 339-347, 2010.
[29] T. Ide, M. A. Hossain, Y. Nakamura, and Y. Ohtsu, Journal of Vacuum Sci. Technol.
A: Vacuum, Surfaces, and Films, 35, 061312, 2017.
108
Chapter 7. Characteristics of a Ring–Shaped Pulsed DC Discharge
Plasma Source Using Single Pole Magnet Setups
7.1. Introduction
Magnetron sputtering (MS) technologies have become the worldwide standard for
most of the physical vapor deposition (PVD) since the invention of the planar cathode
arrangement credited to J. S. Chapin with a patent filed in 1974 [1]. Most of the metallic
materials can be sputtered where oxides and nitrides can be also synthesized. The
depositions show excellent uniformity and reproducibility, smoothed the film surface (no
droplets), and the target (up to a few meters long) can be made with different geometries
in order to coat various shapes of the substrate [2–3]. All these advantages make MS a
highly flexible deposition process and its promising applications are commonly used in
many research, industrial manufacturing sectors and functional coatings preparation [1–
16]. MS glows are, in principle, the sputtering resulting from the target by bombarding
energetic gas ions, where magnetic fields are used to concentrate and confine the plasma
closer to the deposition source [3, 14].
In general, various types of power schemes, such as direct current (DC), radio
frequency (RF), microwave and pulsed DC, have been developed for MS system. Among
these, RF power is mostly used in Plasma Enhanced Chemical Vapor Deposition
(PECVD) and MS system particularly depositing functional films of highly insulating
properties, because the substrates were required to be powered with an alternating pulse
[17-19]. The RFMS method is advantageous to repeat reproducible deposition under an
appropriate impedance matching network between RF power source and powered
electrode load with a matching circuit. For independent control of the RFMS condition,
the substrate bias voltage is not independent of other conditions and depend on the applied
RF power density because of the substrate bias voltage influences incident ion energy on
the deposited functional film [20].
Pulsed DC discharge plasma is considered as an alternative to the RFMS method
and supposed to be a simple and cost-effective deposition system that can be used to
improve film properties as compared to widely used RFMS method [21-23]. In a pulsed
DC discharge plasma source, pulsed DC power supply is a significant component for the
reason that it may possible to change the gas phase chemistry, energy of the incident ions
bombardment on substrate, electron density, electron temperature to meet the
requirements of highly selective deposition method by changing the pulse duration and
109
frequency [24-26]. However, most of the models of commercially available pulsed DC
discharge power supplies are expensive and custom-made which led to less flexibility to
control pulse conditions. A number of researchers reported the pulsed DC discharge
power supplies system mentioning different techniques [27-28]. In comparison with RF
power supplies pulsed DC discharge sources possess a considerably higher efficiency, a
lower cost, a considerably simplified tuning and controlling [20-21]. However, until now
the medium-frequency pulsed discharges are much less studied than radio-frequency or
direct current ones that impede essentially the progress in this area [21, 28].
A large number of theoretical and experimental works have been investigated to
study the pulsed dc discharges plasma [29-41]. However, the problem that the target
utilization rate is not uniform, has been not yet investigated in pulsed DC discharged
plasma. Recently, we have developed radio–frequency ring–shaped magnetized plasma
achieved with magnetic single pole arrangement for uniform target utilization [42]. The
current-voltage characteristics, the electron temperature, the plasma density, the ion and
electron mean free path were analyzed. Moreover, we proposed the outer ring-shaped
radio frequency magnetized plasma source for target utilization in a specific area using
single-pole magnet setups [43]. The relations among various positions of the high-
luminous plasma discharge, peak-magnetic densities and effective ion saturation currents
were discussed in detail for RF power source. The main idea is to analyze the pulsed DC
discharged plasma by the single pole magnet setups at a fixed magnet-gap distance in
order to get high–density ring–shaped plasma. However, the effect of ring–shaped plasma
with single pole magnet setups produced by pulsed DC discharged source had not been
investigated yet in details. Moreover, the differences between RF and pulsed DC
discharges for single-pole magnet setups were not investigated in details.
The objective of this work is to produce a ring–shaped plasma in a specific area
for obtaining outer target erosion profile near the chamber wall. In Section 7.2, the
proposed experimental setup of pulsed DC discharged plasma is explained in detail. In
Section 7.3, the results and discussions on the effects of the three setups are investigated
in a single pole magnet arrangement, three-dimensional magnetic flux lines and their
profiles, gyro–radius of the electrons and ions, Hall parameters of the electrons and ions,
discharge characteristics, ion saturation current profiles are explained. The results
obtained in this work are summarized in section 7.4.
110
7.2. Experimental Arrangements
Figure 7.1 shows a schematic diagram of the proposed pulsed DC discharge
plasma using the single pole magnet setups. The experiments were carried out in a
stainless–steel cylindrical vacuum chamber with an outer diameter of 235 mm, an inner
diameter of 160 mm and a height of 195 mm. The cross-sections of the proposed magnetic
single pole setups of fixed gap distance of 5 mm, including a center magnet, with magnets
in setups (a): one circle, (b): two circles, and (c): three circles are shown in Figs. 7.2(a)–
(c). The pulses were supplied from a pulsed power supply, which is known as pulse
modulator (PuM) to the target. The PuM was powered by three-phase power supply. The
pulse duration of 20μs was set in PuM. The pulse repetition rate was 2.5 kHz. The duty
cycle was 5%. Argon (Ar) gas was introduced into the chamber through a regulating flow
meter at an operating gas pressure of 12 Pa for setups. During the experiments, the target
currents I(t) through the circuit and the target voltage V(t) were measured using a current
monitor and a high–voltage probe, respectively. A current transformer has been used to
measure the instantaneous current at the target. The target was connected to the earth
directly by grounding cable. The consumed electrical power P is calculated by P = V(t)
I(t). A magnetron system equipped with the neodymium cylindrical magnets with a
surface magnetic flux density of 4120 Gauss, diameter of 10 mm, and height of 5 mm
were used. No arc was observed during the experiment. The gap distance between
magnets was fixed at 5 mm in consecutive circles, which are shown in Fig. 7.1. The
number of magnets positioned in setups (a): 1, and 5, (b): 1, 5, and 11, and (c): 1, 5, 11,
and 14. An iron yoke is used to hold the magnets and positioned on the target surface.
The evacuation system consists of a turbo molecular pump (TMP) as a main Evacuator
and a rotary pump (RP) as a mechanical booster pump. Before starting the experiment,
the base pressure in the vacuum chamber of 0.710–5 Pa is attained. The inlet gas pressure
is controlled by a mass flow controller. We selected a fixed gap distance of 5 mm because
high luminescent radio–frequency (RF) magnetized ring–shaped plasmas were attained
with circular magnet monopole setups [42]. Moreover, the ion saturation current
increased significantly for 30 < r < 48 mm [42]. A cooling fan was used to cool the
magnets and the target.
To investigate the pulsed discharge, target voltage and the ion saturation current,
an aluminum target of 140 mm × 140 mm × 3 mm was used, in order to avoid film
deposition on the glass window at the bottom in the chamber. The ion saturation currents,
111
proportional to the ion flux to the target were measured by a negatively biased cylindrical
tungsten Langmuir probe of 1.0 mm in diameter and 10 mm in length [42-43]. The ion
saturation current measurements were done by moving a Langmuir probe of L–shaped in
the radial direction, as shown in Fig. 7.1. The measurement radial positions were set at 0
< r < 48 mm. The biasing voltage was fixed at – 65 V and the voltage drop of the
resistance of 1 k was monitored with a digital oscilloscope. The probe measurements
were carefully performed so as to minimize the disturbance to the stable plasma.
Moreover, the measurement position was fixed at an axial distance, z = 8 mm from the
target surface. The magnetic field is perpendicular to the probe surface in this position so
that electrons are easily collected by the Langmuir probe. As shown in Fig. 7.1, the
origins of the axial distance z and the radial positions, r are defined at the surface and at
the center of the pulsed discharge powered target, respectively.
Figure 7.1. Schematic diagram of the experimental setup used for the proposed pulsed
DC discharge plasma using a single pole magnet setup.
112
113
Figure 7.2 (a)–(c). Proposed single pole magnet arrangements of the ring–shaped
pulsed discharge plasma source including a center magnet, with magnets in setups (a):
one circle, (b): two circles, and (c): three circles for material processing.
7.3. Results and Discussions
7.3.1. Simulation of Single Pole Magnet Setups for Ring-Shaped Pulsed Discharge
The three-dimensional (3D) magnetic flux lines for the three magnet setups (a),
(b), and (c) are analyzed using Femtet software in order to discuss the effect of the
magnetic field profiles in the proposed single pole magnet setups on the pulsed discharge.
Figure 7. 3 shows a map of the three-dimensional distribution of the magnetic flux lines
generated by the permanent magnets on the target surface, z = 0 mm for setups (a), (b),
and (c). A circular iron disk, that is, magnet holder of 140 mm of diameter and 1 mm of
thickness diameter was used. The magnetic field lines generated from the N-pole of the
magnets passed through the target, the airspace, the magnet holder and then returned to
the S–pole for all single–pole magnet arrangements as shown in Fig. 7. 3(a)–(c). The
axisymmetric property for the single–pole magnet setup is satisfied because the flux lines
profile for the half region from r = −70 to 0 mm is similar to that for the half region from
114
r = 0 to +70 mm [see Figs 7.3 and 7.4]. It is important to effectively use the E × Bx drift
motion, to generate ring–shaped pulsed discharge at the specific areas, where E and Bx
are the electric field perpendicular to the target and the magnetic flux density parallel to
the target, respectively. Moreover, the simulated magnetic flux density component
parallel to the target has a significant effect on the E × Bx drift motion, the gyro–radius,
and the Hall parameter. Setup (c) has high-density magnetic flux lines distributions in the
outer area than setups (a) and (b), where setup (a) has the lowest among three setups from
approximately r = 20 to 50 mm [Fig. 7. 3(a)–(c)]. The highest 2D magnetic flux lines
concentrations were found approximately at 18 to 23 mm, 33 to 38 mm, and 48 to 53 mm
for setups (a), (b), and (c), respectively.
115
Figure 7. 3 (a)–(c). Maps of the two-dimensional distribution of the magnetic flux lines
generated by the permanent magnets on the target surface for setups (a), (b), and (c).
116
Figure 7.4 shows the radial profile of the absolute value of (a) the horizontal
magnetic flux density, |Bx| and (b) the vertical magnetic flux density |Bz| for the setups (a),
(b), and (c) at an axial distance of z = 0 mm from the target surface. Multiple peaks are
obtained in all magnet arrangements. The patterns of the |Bx| for setup (b) is almost same
as setup (c) from r = −23 to 23 mm. The peaks of the horizontal magnetic flux density,
|Bx| profiles are 6, 10, and 14 for the setups (a), (b), and (c), respectively. The magnetic
flux density from r = 0 to ±45 mm for setup (c) had the lowest magnitudes among the
three setups. A maximum flux density of approximately 521 G at r = 0 to ±16 mm for
setup (a), 496 G at r = 0 to ±31 mm for setup (b), and 504 G at r = 0 to ±46 mm for setup
(c) were observed with a fluctuation from zero. Moreover, the three highest peak values
were 1080 G at r = ±19 mm, 1166 G at r = ±36 mm, 1158 G at r = ±50 mm for the setups
(a), (b), and (c), respectively. The peak amplitude of |Bz| for setup (a) is lower than that
for setups (b) and (c) from r = 0 to ±20 mm. The peak amplitudes of |Bz| approximately
846 G, 1000 G, and 373 G at r = 0, ±15, and ±30 mm for setup (a), approximately 588 G,
642 G, 884 G, and 444 G at r = 0, ±15, ±30, and ±45 mm for setup (b), approximately
474 G, 493 G, 574 G, and 816 G at r = 0, ±15, ±30, and ±45 mm for setup (c) were found,
respectively. It is noted that the vertical magnetic flux density components do not
contribute to the E × B drift motion of the charged particles.
(a)
117
(b)
Figure 7.4. Radial profiles of the absolute value of (a) the horizontal, |Bx| and (b) the
vertical, |Bz| magnetic flux density at the target surface for setups (a), (b), and (c).
The cyclotron angular frequency of the charged particles such as electrons and
ions are important parameters for magnetized pulsed discharge plasma. These calculated
cyclotron frequencies based on the magnetic field density, Bx, the electronic charges, e,
and the masses of charged particles, m, were 1.9×1010, 2.1×1010, and 2.0×1010 Hz at a
high flux density position for setups (a), (b), and (c), respectively. These high–magnitudes
flux density will provide some ring–shaped discharges due to the E×B drift effect in the
specific areas for all setups. The larmor radius is also known as gyro radius. The larmor
radius is investigated to discuss the magnetization of the charged particles in the
magnetized pulsed discharge plasma. The larmor radius of the electrons is calculated
from , where m, v, e, and Bx denote the charged particle mass, the velocity of
the charged particles, the electronic charge, and the magnetic flux density in the radial
direction, respectively.
Figure 7.5 shows radial profiles of the larmor radius of the (a) electrons and (b)
ions, respectively. The larmor radius is also known as gyro radius. The larmor–radii are
calculated by the magnetic flux density in the radial direction at an axial position of z =
0 mm with the assumption of electron temperature, Te = 2 eV, ion temperature of Ti = 0.2
eV, and the charged particle velocity of 9.5×105 ms-1 [43]. It is seen that the electron
larmor radius, rLe are less than 4 mm except at r = 0 mm position for all setups. For
example, the electron velocity, ve = 9.5 × 105 ms-1 assumed by the electron temperature
118
Te = 2 eV, at r = 40 mm, z = 0 mm, and the magnetic flux density in the radial component
(a): Bx = 158 G, (b): Bx = 803 G, (c): Bx = 295 G, the estimated electron larmor radius, rLe
is approximately (a): 0.34 mm, (b): 0.06 mm, and (c): 0.18 mm, respectively. Moreover,
the electron larmor radius rLe are 0.17 mm, 1.64 mm, 5.82 mm for setups (a), (b), and (c),
respectively. In single pole magnet arrangements, the highest value for the electron gyro–
radius is approximately (a): 17.23 mm, (b): 9.67 mm, and (c): 9.40 mm, at r = 0 mm,
where the magnetic flux density in the radial direction has the lowest value. It is also seen
that the electron gyro–radius is much smaller than the size of the chamber radius. On the
other hand, the argon ion larmor radius, rLi has the maximum value of 108m, 17.76m,
17.27m at r = 0 mm for setups (a), (b), and (c), respectively. The radial profile of argon
ion larmor radius is the same as the profile of the electron larmor radius. Moreover, argon
ion larmor radius is higher than the electron larmor radius corresponding to each point.
(a)
119
(b)
Figure 7.5. Radial profiles of the larmor radius of the (a) electrons and (b) ions for
setups (a), (b), and (c).
Figures 7.6(a)–(b) show the Hall parameters (HP) of electrons and ions at z = 0
mm for setups (a), (b), and (c) in a logarithmic scale. The Hall parameter for electrons
indicates the magnetization degree, which represents a rotating number of electrons
around the magnetic field in their collision times. The HP is estimated by the product
between the electron cyclotron angular frequency, ωc and the electron collision time, τce
which is inversely proportional to the electron collision frequency, ʋme [17, 43]. The
electron Hall parameter, he and the ion Hall parameter, hi of electrons are expressed as he
= ωce/ʋmi and hi = ωci/ʋmi where ωce, ωci, ʋme, and ʋmi are the cyclotron angular frequencies
of electrons and ions, and collision times of electrons and ions, respectively.
It is seen that the radial profiles of the Hall parameters are inversely proportional
to the radial profiles of the gyro-radii [see Figs. 7.5 and 7.6]. This is reasonable because
the cyclotron frequency is the reciprocal of the gyro–radius. The highest electron Hall
parameters of approximately 561 at r = ±21 mm for setup (a), 544 at r = ±36 mm for
setup (b), and 297 at r = ±50 mm for setup (c) were found. The electron Hall parameters
120
are much larger than one for all magnet setups except r = 0 and ±16 mm for setup (a), r
= 0 and ±31 mm for setup (b), and r = 0 and ±46 mm for setup (c), respectively. Thus,
the electrons in the pulsed discharges are strongly magnetized at the regions with high
electron HP. Thus, the energetic beam electrons are accelerated in the pulsed discharges
by single–pole magnet setups. So that along the magnetic field line, the bounce resonance
heating happens to produce the ring–shaped pulsed discharges. The highest ion Hall
parameters of approximately 0.3 at r = ±21 mm, 0.29 at r = ±36 mm, and 0.16 at r = ±50
mm were found for setups (a), (b), and (c), respectively as shown in Fig. 7.6 (b).
Moreover, all the ions Hall parameters are much less than one, whereas the radial profile
has the same pattern as the electron Hall parameter. Therefore, all ions are not magnetized
because all the ion gyro–radius in the pulsed discharges are more than the size of the
vacuum chamber radius.
(a)
121
(b)
Figure 7.6. Radial profiles of the Hall parameters of the (a) electrons and (b) ions for
setups (a), (b), and (c).
7.3.2. Electrical Characteristics of the Ring–Shaped Pulsed Discharge Plasma
Figure 7.7 shows the temporal behavior waveforms of (a) target voltage, (b)
current through the target (the glow current), and (c) power consumed in the plasma, at
Ar gas pressure of 12 Pa, the pulse repetition rate of 2.5 kHz for setups (a), (b), and (c),
respectively. The substrate voltage was fixed at –65V. The source voltage applied to the
target is 0.6 kV, 0.6 kV, and 1.1 kV for setups (a), (b), and (c), respectively and its pulse
has a duration of 20μs. Longer pulses were not applied due to severe arcing. The voltage
applied after 10 μs have passed and then the ionization started. The peak currents were
approximately 0.14 A, 0.38 A, and 0.28 A for setups (a), (b), and (c), respectively. It
seems that the current is very small. This is reasonable because the output impedance of
the pulse modulator was 1.5 kΩ. At the end of the pulse, the current through the target
becomes lower (almost zero). Moreover, the current through the target for setup (c) is
about –0.26 A due to the displacement current and voltage changes to 0 V very rapidly
after the end of the pulse. The peak powers consumed in the plasma for pulsed discharges
were approximately 72.8 W, 187.2 W, and 171.6 W for setups (a), (b), and (c),
122
respectively. The consumed power in the plasma is also low because the current through
the target is limited by the output impedance of 1.5 kΩ of the pulse modulator.
(a)
(b)
123
(c)
Figure 7.7. Waveforms of (a) target voltage, (b) current through the target, and (c)
power consumed in the plasma at Ar gas pressure of 12 Pa, pulse repetition rate of 2.5
kHz for setups (a), (b), and (c). The output impedance of pulse modulator was 1.5 kΩ.
7.3.3. Discharge Characteristics of the Ring–shaped Pulsed Discharge Plasma
Figures 7.8(a)–(c) show the typical photographs of the pulsed plasma DC
discharge plasma at Ar gas pressure of 12 Pa, pulse duration of 20μs, and a pulse
repetition rate of 2.5 kHz for setups (a), (b), and (c), respectively. The white dashed circle
denotes the position of the outer wall of the chamber. The pulsed DC discharge
magnetized plasma photographs were taken using a low-resolution digital camera. The E
B drift motion of the electrons plays a significant effect on the distribution of plasma
discharge as mentioned in the above section. The effect of high plasma density and the
superposition of the electric and the magnetic field results in the reformation of the dense
ionization regions of the target in pulsed DC discharge plasma, which is good agreement
with Y. Ohtsu et al [44]. These ionization regions were captured on the target surface and
also observed to intensify and extend along the target racetrack with increasing Ar
pressure from 6 Pa to 12 Pa. The strong ring–shaped plasma discharge is observed for all
setups. Multiple ring–shaped plasma discharges were not found for all setups. The high
luminous plasma positions were approximately at 20 mm, 40 mm, and 48 mm, where
almost correspond to the positions of the highest electron HP, for setups (a), (b), and (c),
respectively. The typical discharge voltages were 0.6 kV, 0.6 kV, and 1.0 kV for setups
(a), (b), and (c), respectively.
124
(a)
(b)
125
(c)
Figure 7.8. Typical photographs of the pulsed DC discharge plasma at Ar gas pressure
of 12 Pa for setups (a), (b), and (c). The dashed circle denotes the outer wall of the
chamber.
Figure 7.9 shows the target voltage and duty cycle as a function of pulse
repetition rate for setup (b). It was difficult to do experiments for setups (a) at 0.7 kV or
more, (b) at 0.7 kV or more, and (c) at 1.1 kV or more because the overloading currents
trip the circuit breaker and shut down the pulsed power source. The setups (a) and (b)
were easy for generating pulsed discharge plasma compared to setup (c). Moreover, the
critical discharge point for the proposed pulsed discharge was also observed. There was
no pulsed discharge plasma at ≤ 2 kHz. . Moreover, plasma discharges were not observed
at low gas pressure less than 6.0, 6.0, and 12.0 Pa for setups (a), (b), and (c), respectively.
126
Figure 7.9. Target voltage and duty cycle as a function of pulse repetition rate for setups
(b) to investigate the critical point for the proposed pulsed discharge plasma.
7.3.4. Ion Saturation Currents Profiles
Figure 7.10 shows the temporal waveforms of the ion saturation currents
detected by a tiny Langmuir probe positioned at z = 8 mm from the target surface, the
radial position of r = 40 mm, Ar gas pressure of 12 Pa current, a fixed bias voltage of –
65 V, pulse duration of 20 μs and the pulse repetition rate of 2.5 kHz for setups (a), (b),
and (c). Ion saturation current, Iisat can be used as a proxy for the incoming ion flux to the
target, which in turns has a significant effect on the film properties. The ion saturation
current, Iisat for setup (c) dominates for the first 20 μs of the pulse, while later in the pulse
the Iisat for setup (b) dominates. This is because the application voltage was high for t <
20 μs for setup (c). At the end (0 μs) of the pulse, the target voltage drops instantaneously
to zero [see figure 7(a)], and the ions are not returned immediately by the high target
voltage after the pulse. As a result, Iisat for all setups gradually decreases. The Iisat were
3.2 mA, 2.8 mA, and 4.3 mA at t = –10μs for setup (a), (b), and (c), respectively. A
resistance of 1.5 k was used as the output impedance of the pulse modulator.
127
Figure 7.10. Temporal behavior of the ion saturation currents as a function of time at
the radial position of r = 40 mm, Ar pressure of 12 Pa current, and z = 8 mm for setups
(a), (b), and (c). Here, the lower value of peak Iisat observed because the output
impedance of pulse modulator was 1.5 kΩ.
Figure 7.11 shows the ion saturation current, Iisat of the pulsed DC discharge as
a function of radial position at a fixed time (t = –12 μs, –5 μs, +10 μs, +40 μs), at Ar gas
pressure of 12 Pa for setups (a), (b), and (c). The ion saturation currents for setups (a),
(b), and (c) were approximately same for the radial position of r = 0 to 20 mm at t = –12
μs, –5 μs, +10 μs. The highest ion saturation currents were found at t = –5 μs for all setups.
The Iisat of setup (b) is less fluctuating, whereas Iisat of setup (c) is highly variable in all
radial position at t = –12 μs, –5 μs, +10 μs. It is important to note that, the highest ion
saturation currents were found for setup (b) at a radial position of r = 45 mm at all fixed
time. The Iisat of 1.44, 2.88, and 2.2 mA for setups (a), (b), and (c), respectively at r = 45
mm and t = +10 μs were found. It is seen that, as the time passes, the ion saturation
magnitudes are decreasing gradually [see Figs. 7.11 (a)–(d)]. This effect is reasonable
due to plasma diffusion. The decay of the ionization is occurred by electron localization
in a ring–shaped pulsed plasma discharge using single pole magnet setups. The lowest
Iisat is observed at t = +40 μs for all setups because the target voltage and current through
the plasma are returned to zero. Moreover, ion saturation current decreases very slowly
near the chamber wall for setups (b) and (c).
128
(a)
(b)
(c)
129
(d)
Figure 7.11. Radial profiles of the ion saturation currents of the pulsed discharge at Ar
pressure of 12 Pa at t = –12 μs, –5 μs, +10 μs, and +40 μs for setups (a), (b), and (c).
7.4. Conclusion
The proposed single pole setups of fixed gap distance of 5 mm, including a
center magnet, with magnets in setups (a): one circle, (b): two circles, and (c): three
circles were used. The number of magnets placed in setups (a): 1, and 5, (b): 1, 5, and 11,
and (c): 1, 5, 11, and 14. The pulse duration of 20μs, the pulse repetition rate was 2.5 kHz,
Ar pressure of 12 Pa current, and the source voltage to the target is (a): 0.6 kV, (b): 0.6
kV, and (c): 1.1 kV were applied to do pulsed discharge experiments. The highest two
dimensional magnetic flux lines concentrations were found approximately at (a): 18 to
23 mm, (b): 33 to 38 mm, and (c): 48 to 53 mm. The peaks of the horizontal magnetic
flux density, |Bx| profiles were 6, 10, and 14 for the setups (a), (b), and (c), respectively.
The three highest peak values of |Bx| were 1080 G at r = ±19 mm, 1166 G at r = ±36 mm,
1158 G at r = ±50 mm for the setups (a), (b), and (c), respectively. It is also seen that the
electron larmor radius is much smaller than the size of the chamber radius. The argon ion
larmor radius has the maximum value of 108 m, 17.76 m, 17.27 m at r = 0 mm for setups
(a), (b), and (c), respectively. The electron Hall parameters are much larger than one for
all magnet setups except r = 0 and ±16 mm for setup (a), r = 0 and ±31 mm for setup (b),
and r = 0 and ±46 mm for setup (c), respectively, whereas all the ions Hall parameters
are much less than one. The peak currents in the plasma were approximately 0.14 A, 0.38
A, and 0.28 A for setups (a), (b), and (c), respectively, whereas the output impedance of
the pulse modulator was 1.5 kΩ. The high luminous plasma discharge positions were
130
approximately at 20 mm, 40 mm, and 48 mm for setups (a), (b), and (c), respectively.
The ion saturation current, Iisat for setup (c) dominates for the first 20 μs of the pulse and
the Iisat for setup (b) dominates at t >20 μs. The highest ion saturation currents were found
for setup (b) at a radial position of r = 45 mm at all fixed time. Later, functional films
will be prepared using this plasma source.
131
References
[1] J. S. Chapin, Sputtering Process and Apparatus, U.S. Patent 4, 166, 018, August 28,
1979 (Submitted 1/31/1974).
[2] A. Caillard, M. El. Mokh, T. Lecas, A. L. Thomann, Vacuum, 147, 82-91, 2018.
[3] G. Brauer, B. Szyszka, M. Vergohl, and R. Bandorf, Vacuum, 84, 1354-1359, 2010.
[4] F. Fietzke and O. Zywitzki, Thin Solid Films, 644, 138–145, 2017.
[5] Y. Ohtsu, N. Matsumoto, J. Schulze, and E. Schuengel, Phys. Plasmas, 23, 033510,
2016.
[6] E. Schungel, R. Hofmann, S. Mohr, J. Schulze, J. Ropcke, and U. Czarnetzki, Thin
Solid Films, 574, 60–65, 2015.
[7] M. A. Hossain, T. Ide, K. Ikari, and Y. Ohtsu, Vacuum, 128, 219-225, 2016.
[8] T. Lafleur and R. W. Boswell, Phys. Plasma, 19, 2, 023508-1–023508-12, 2012.
[9] T. Ide, M. A. Hossain, Y. Nakamura, Y. Ohtsu, Journal of Vacuum Sci. Technol. A:
Vacuum, Surfaces, and Films, 35, 061312, 2017.
[10] Q. Guo, W. Shi, F. Liu, T. Nakao, K. Saito, T. Tanaka, and M. Nishio, Jpn J. of
Applied Physics, 51,118004, 2012.
[11] S. Nakao, K. Yukimura, S. Nakano, and H. Ogiso, IEEE Trans. on Plasma Sci., 41,
8, 2013.
[12] U. Helmersson, M. Lattemann, J. Bohlmark, A. P. Ehiasarian, and J. T.
Gudmundsson, Thin Solid Films, 513, 1–2, 1–24, 2006.
[13] H. Takikawa, K. Izumi, R. Miyano, and T. Sakakibara, Surface Coating Technol.,
163–164, 368–373, 2003.
[14] S. Spagnolo, M. Zuin, R. Cavazzana, E.Martines, A. Patelli, M. Spolaore and M.
Colasuonno, Plasma Sources Sci. Technol. 25, 065016, 2016.
[15] Z. Chen, K. Nishihagi, X. Wang, C. Hu, M. Arita, K. Saito, T. Tanaka, and Q. Guo,
Luminescence, 194, 374–378, 2018.
[16] X. Wang, Z. Chen, C. Hu, K. Saito, T. Tanaka, M. Nishio, and Qixin Guo, Crystal
Growth, 483, 39–43, 2018.
[17] M. A. Lieberman, and A. J. Lichtenberg, “Principles of plasma discharges and
material processing”, 2nd edn. John Wiley & Sons, Inc., New York, 2005.
[18] C. Corbella, M. Vives, G. Oncins, C. Canal, J.L. Andújar, E. Bertran, Diamond
Related Materials, 13, 4, 1494–1499, 2004.
132
[19] F. F. Chen, and J. P. Chang, “Lecture notes on principles of plasma processing”,
Kluwer Academic, New York, 2003.
[20] M. A. A. Mamun, H. Furuta, and A. Hatta, Materials Today Communica., 14, 40-46,
2018.
[21] V. A. Lisovskiy, P. A. Ogloblina, S. V. Dudin, V. D. Yegorenkov, and A. N. Dakhov,
Vacuum, 145, 194-202, 2017.
[22] R.A. Scholl, Surface Coating Technol., 98, 823, 1998.
[23] V. J. Trava-Airoldi, L. F. Bonetti, G. Capote, J. A. Fernandes, E. Blando, R. Hubler,
P. A. Radi, L. V. Santos, and E. J. Corat, Thin Solid Films, 516, 272-276, 2007.
[24] G. Lotito, T. Nelis, P. Guillot, and D. Gunther, Spectrochimica Acta B, 66, 619–626,
2011.
[25] S. Samukawa, and T. Mieno, Plasma Sources Sci. Technol., 5, 2, 132, 1996.
[26] M.A. Lieberman, and S. Ashida, Plasma Sources Sci. Technol., 5, 2, 145, 1996.
[27] J. Choi, Electrical Engineering Technol. 5, 3, 484–492, 2010.
[28] W. Jiang, Proc. IEEE, 92 (7), pp. 1180–1196, 2004.
[29] F. Clement, B. Held, N. Soulem, N. Spyrou, Eur. Phys. J. AP, 13, 1, 67-73, 2001.
[30] F. Clement, B. Held, N. Soulem, Eur. Phys. J. AP, 17 2, 119-130, 2002.
[31] E.V. Barnat, T.M. Lu, Phys. Rev. E: Statistical, Nonlinear, and Soft Matter Physics,
66, 056401, 2002.
[32] H. Coitout, G. Cernogora, J. Phys. D. Appl. Phys., 39, 1821, 2006.
[33] S. Gershman, O. Mozgina, A. Belkind, K. Becker, E. Kunhardt, Contributions
Plasma Phys., 47, 19-25, 2007.
[34] P.K. Chu, and X.P. Lu, “Low Temperature Plasma Technology: Methods and
Applications”, CRC Press, Boca Raton, FL, 2014.
[35] X.B. Tian, P.K. Chu, Phys. D: Appl. Phys., 34, 354, 2001.
[36] L.M. Isola, B.J. Gomez, V. Guerra, Phys. D: Appl. Phys., 43, 015202, 2010.
[37] S. Filimonov, J. Borysow, Phys. D: Appl. Phys., 40, 2810, 2007.
[38] C.D. Pintassilgo, V. Guerra, O. Guaitella, A. Rousseau, Plasma Sources Sci.
Technol., 19, 055001, 2010.
[38] C.D. Pintassilgo, V. Guerra, O. Guaitella, A. Rousseau, Plasma Sources Sci.
Technol., 23, 025006, 2014.
[40] V. Efimova, “Study in Analytical Glow Discharge Spectrometry and its Application
in Materials Science”, PhD Thesis, Technische Universitat Dresden, Dresden, 2011.
133
[41] V. Efimova, V. Hoffmann, J. Eckert, Analytical Atomic. Spectrometry, 26, 784-791,
2011.
[42] M. A. Hossain and Y. Ohtsu, Jpn. J. Appl. Phys., vol. 57, 01AA05, 2018.
[43] M. A. Hossain and Y. Ohtsu, IEEE Transactions on Plasma Science, vol. 46, no. 8,
August 2018.
[44] Y. Ohtsu and T. Yanagise, Plasma Sources Sci. Technol., 24, 034005, 2015.
134
Chapter 8. Summary and Suggestions for Future Plan
8.1. Summary and Conclusions
Chapter 1 discussed the plasma and its typical features, thin film deposition by
plasma sputtering background research on plasma sputtering problems of the plasma
sputtering, objectives of this Thesis, and structure of this thesis. Different measurement
techniques and methods of the ion saturation currents, plasma density, and electron
temperature, eroded target profile and target utilization rate calculation, resistivity and
thickness of the film were explained. In the chapter 3, a square-shaped magnet
arrangement consisting of eight neodymium rod magnets of 30 × 5 × 3 mm has been
investigated in order to realize a uniform utilization of a copper target in an RF
magnetized sputtering plasma. A copper plate of 160 × 160 × 3 mm and an aluminum
disc of 160 mm in diameter were used as the target to measure the sputtering
characteristics based on rotational square-shaped magnet schemes in the high-density RF
magnetized plasma sputtering source. The magnetic shielding material, iron (Fe) cover,
suppresses the horizontal magnetic flux density |Bx| at x = 0 from approximately 65 to 50
mT. Strong plasma emission in an outward direction is observed at four intersections
between the four vertically placed magnets. It is seen that the RF discharge voltage and
the absolute value of the dc self-bias voltage decrease gradually from approximately 760
to 600 Vp-p and approximately 400 to 270 V, respectively, with increasing argon gas
pressure for all cases. The uniformity of the target erosion depth and ion flux profile has
been improved by using the iron (Fe) cover and decreasing the air gap between the
shielded magnets. The iron (Fe) cover is effective to prohibit the inward E×B drift motion
to the center of the target so that the uniformity of the copper erosion profile is improved
by the addition of the iron cover and reducing the air gap between the one–pair shielded
magnet and the target utilization reaches the highest value of 87.49%.
In the chapter 4, a copper film was deposited using a gyratory square-shaped
magnet arrangement consisting of eight neodymium bar magnets of 30 mm × 5 mm × 3
mm. Two magnetic arrangements were investigated to discuss the film properties: case
(a) without iron shielding and case (b) with iron shielding of 5 mm × 3 mm × 1 mm,
which is used for magnetic shielding. Copper films were prepared at an RF power of 100
W, an Ar pressure of 1.0 Pa, a processing time of 1.5 h, a gyratory speed of 40 rpm, and
a target– substrate distance of 45 mm. The magnetic shielding material reduces the
horizontal magnetic flux density, |Bx|, at x = 0 from ~1120–824 G and the axial
135
component, |Bz|, from ~824–706 G at x = -15 and 15 mm. The RF discharge voltages,
VRF, (peak-to-peak value) were 1080 V and 1160 V, and the absolute values of the self-
bias dc voltages, VSB, of the electrode were 528 V and 573 V for cases (a) and (b),
respectively. A typical star-shaped plasma discharge has been observed. The Cu films for
case (a) exhibit sharp needle-shaped grains vertical to the substrate with a mean grain size
of ~40 nm, and for case (b) the film surface shows an improved smooth surface with a
reduced mean grain size of ~20 nm. The roughnesses of the copper thin films are ~3.73
nm and ~2.49 nm for case (a) and case (b), respectively. An XRD investigation shows
that the FWHM values of the diffraction peak were 0.42° and 0.66° for cases (a) and (b),
respectively and both samples exhibit a face-centered-cubic structure.
In the chapter 5, an outer ring-shaped RF magnetized plasma has been proposed
by three circular monopole magnet arrangement such for (a) R = 5 mm, (b) R = 20 mm
and (c) R = 35 mm It is seen that a high luminous ring-shaped plasma is observed in
contiguity with the wall position of r = 47-50 mm for (a) R = 5 mm, whereas multi-ring
discharges are observed for (b) R = 20 and (c) 35 mm. The tendency of the absolute value
of the DC self-biased voltage |Vsb| of the cathode is almost similar to the pattern in the
discharge voltage |Vrf| corresponding to the three monopole schemes. It is found that the
electron temperature decreases with increasing gas pressure for all cases. However, it is
seen that the plasma density increases with increasing gas pressure for all setups. The
smaller gap distance, the fluctuations of the plasma density and ion saturation current are
more. However, a big gap distance as (c): R = 35 mm, has high ion saturation current with
more fluctuations comparison with case (b): R = 20 mm. So that case (b): R = 20 mm is
effective to improve target utilization from center to the outer area. From a viewpoint of
the center to outer target utilization, case (b): R =20 mm is more effective than other two
cases. However, case (a): R = 5 mm can be used to improve target utilization only in the
outer area, whereas (c): R = 35 mm has the highest ion saturation current and lowest
electron temperature among all cases.
In the chapter 6, a ring-shaped plasma in the outer area is observed at the position
of the peak magnetic flux density found in monopole magnet arrangement and its
diameter depends on a number of magnet circles. The high-density plasma was found in
the outer target region near the chamber wall for all type monopole arrangements. The
plasma discharge results that the target utilization can be controlled effectively in the
outer specific area near the chamber wall. The electron hall parameter was (a): 245, (b):
136
395, and (c): 393 at high luminescent plasma discharge position. It was observed that the
ion saturation current, Iisat are 0.6 mA, 0.79 mA, and 0.46 mA, for setups (a), (b), and (c),
respectively at a radial position of r = 47 mm. The Iisat for setup (a) is the lowest for 0 <
r < 35 mm, whereas for setup (c), Iisat has the highest value among three setups. It is seen
that, Iisat is very high in the outer target region near the chamber wall for setups (a) and
(b), where Iisat for setup (c) decreases slowly. Thus, the proposed CCP plasma source can
be used for the outer target erosion and target utilization in the specific area.
In the chapter 7, A pulsed DC plasma discharge is proposed. The pulse duration
of 20μs, the pulse repetition rate was 2.5 kHz, Ar pressure of 12 Pa current, and the source
voltage to the target is (a): 0.6 kV, (b): 0.6 kV, and (c): 1.1 kV were applied in pulsed DC
discharge plasma source experiments. It was found that the three highest peak values of
|Bx| were 1080 G at r = ±19 mm, 1166 G at r = ±36 mm, 1158 G at r = ±50 mm for the
setups (a), (b), and (c), respectively. It is also seen that the electron larmor radius is much
smaller than the size of the chamber radius, while the argon ion larmor radius has the
maximum value of 108 m, 17.76 m, 17.27 m at r = 0 mm for setups (a), (b), and (c),
respectively. The electron Hall parameters are much larger than one for all magnet setups
except r = 0 and ±16 mm for setup (a), r = 0 and ±31 mm for setup (b), and r = 0 and ±46
mm for setup (c), respectively, whereas all the ions Hall parameters are much less than
one. The peak currents in the plasma were approximately 0.14 A, 0.38 A, and 0.28 A for
setups (a), (b), and (c), respectively, whereas the output impedance of the pulse modulator
was 1.5 kΩ. The high luminous plasma discharge positions were approximately at 20 mm,
40 mm, and 48 mm for setups (a), (b), and (c), respectively. The ion saturation current,
Iisat for setup (c) dominates for the first 20 μs of the pulse and the Iisat for setup (b)
dominates at t >20 μs. The highest ion saturation currents were found for setup (b) at a
radial position of r = 45 mm at all fixed time.
137
8.2. Suggestions for Future Works
Based on the present study, the results could be improved via some suggestions.
A comparison may be summarized based on the experimental works between RF
magnetized plasma and Pulsed DC discharge using single-pole magnet setups. Few new
magnet setups can be investigated experimentally to get the target utilization of
approximately 9095% Moreover, Ring-shaped plasma for specific area can be used for
different kind of functional films preparation and the film characteristic may be
summarized. For next step, it is recommended that various functional films can be
deposited by the developed square-shaped and monopole setups plasma source. The
evaluation of the prepared functional film will be analyzed by some useful technical tool
such as surface profiler, four-point probe method, atomic force microscopy (AFM),
scanning force microscopy (SEM), raman spectroscopy, x-ray photoelectron
spectroscopy (SPS), x-ray diffraction (XRD), optical emission spectrometer (OES),
Ultraviolet Visible Near-Infra red (UV-Vis-NIR) spectroscopy and Hall measurement
techniques.
138
List of Publications
International Journals
(1) M. A. Hossain, T. Ide, K. Ikari, and Y. Ohtsu, “High-density radio-frequency
magnetized plasma sputtering source with rotational square-shaped arrangement
of rod magnets for uniform target utilization,” Vacuum, vol. 128, pp. 219-225,
Jun. 2016.
(2) M. A. Hossain, Y. Ohtsu, and T. Tabaru, “Performance of a Gyratory Square-
Shaped Capacitive Radio Frequency Discharge Plasma Sputtering Source for
Materials Processing,” Plasma Chem. Plasma Process, vol. 37, pp. 1663–1677,
2017.
(3) Md. Amzad. Hossain and Yasunori Ohtsu, “RF magnetized ring-shaped plasma
for target utilization obtained with circular magnet monopole arrangement,” Jpn.
J. Appl. Phys., vol. 57, 01AA05, 2018.
(4) Md. Amzad. Hossain and Yasunori Ohtsu, “Outer Ring-Shaped Radio
Frequency Magnetized Plasma Source for Target Utilization in Specific Area”,
IEEE Transaction on Plasma Science, Manuscript accepted dated on 14 June,
2018.
(5) M. A. Hossain and Yasunori Ohtsu, Tentative title “Observation of Ring-shaped
magnetized pulsed DC discharge plasma source using single pole magnet setups”,
Prepared manuscript will be submitted in refereed international journal on middle
of August 2018.
139
Conference Proceedings
(1) M. A. Hossain and Y. Ohtsu, “Development of cross-shaped magnetized plasma
by square-shaped magnets and uniform utilization of Cu target”, JSAP annual
spring meeting, March 2016, Tokyo, Japan.
(2) M. A. Hossain and Y. Ohtsu, “Electrical and Structural Properties of Copper Thin
Films Deposited by Novel RF Magnetized Plasma Sputtering with Gyratory
Square-Shaped Arrangement by Bar Permanent Magnets”, 69th Gaseous
Electronics Conference (GEC 2016), October 2016, Bochum, Germany.
(3) M. A. Hossain and Y. Ohtsu, “RF Magnetized Ring-Shaped Plasma for Uniform
Target Utilization by Circular Magnets Monopole Arrangement”, accepted in 9th
International Symposium on Advanced Plasma Science and its Applications for
Nitrides and Nanomaterials/10th International Conference on Plasma-Nano
Technology & Science (ISPlasma2017/IC-PLANTS2017), Nagoya, Japan.
(4) M. A. Hossain and Y. Ohtsu, “Outer Circular Ring-Shaped RF Magnetized
Plasma for Specific Area Target Utilization by Magnetic Monopole
Arrangement”, 44th IEEE International Conference on Plasma Science (ICOPS
2017), May 2017, Atlantic City, New Jersey, USA.
(5) M. A. Hossain and Y. Ohtsu, “Outer Ring-Shaped Magnetized Plasma by RF and
HiPIMS source”, 70th Gaseous Electronics Conference (GEC 2017), November,
2017, Pittsburgh, Pennsylvania, USA.
(6) M. A. Hossain and Y. Ohtsu, “Ring-Shaped Plasma for Target Utilization in
Specific Area by HiPIMS Source”, 39th International Symposium on Dry Process
(DPS2017), November, 2017, Tokyo, Japan.
(7) M. A. Hossain, Y. Nakamura, A. Sugawara, and Y. Ohtsu, “Preparation of
aluminum doped zinc oxide film by radio-frequency magnetized plasma
sputtering source with square-shaped rod magnets”, 19th International Congress
on Plasma Physics (ICPP 2018), June 2018, Vancouver, Canada.